Digipix3 v300 en

!
Compendium lor digilal pholography

Image
Engineering
English translation sponsored by:
COMPENDIUM FOR DIGITAL PHOTOGRAPHY
A cooperation of :
Compendium lor digilal pholography 2
Image
Engineering
Dear readers.
The "Compendium for digital photography" aka. "DigiPix" in version
3 was originally written and published November 2004 in and by the
cooperation of the adf (Working Group Digital Photography) and the
ECI (European Color Initiative) - in german only. With the kind of posi-
tive acceptance we received in public the idea of an international ver-
sion available in english was soon born and eventually realized. Due
to the fast paced innovation in the feld of digital photography and
digital imaging in general some areas of this guide are not up-to-date
with the latest and greatest. An updated version will be available in
the near future.
Meanwhile the authoring team hopes to provide you, the photog-
rapher with a batch of tips and helpful background info, in order to
make your daily digital job a lot easier.
I would like to thank the sponsors, Image Engineering Dietmar Wller,
Tecco Braun & Frings GmbH and LaserSoft Imaging AG for making this
translation possible through their much appreciated support.
We wanted to make sure that this publication can be used as freely as
possible which is why we issued CREATIVE COMMONS PUBLIC LICENSE
at the end of this document. We the authors expect the compliance
with it.
Jan-Willem Rosse September of 2006
(jan@rossee.de)
The authors:
adf (workingroup digital photography):
Reinhard Fittkau, Holger Hagedorn, Dirk Hartmann, Dr. Martin Knapp,
Matthias Weise
ECI:
Dieter Dolezal, Dietmar Fuchs, Jan-Willem Rosse, Andre Schtzen-
hofer, Dietmar Wueller
weitere Autoren:
Ulrike Haessler, Christian Westphalen
Preambel
3 Compendium lor digilal pholography
Image
Engineering
Preamble 2
Content 3
1. Color management 4
1.1 Basics color management 4
1.2 Profling a digital camera 8
1.3 Working spaces 13
1.4 Viewing conditions 18
1.5 Visualization on the monitor 20
1.6 Visualization in print 24
1.7 Settings in Photoshop 27
2 Photography 36
2.1 Exposure and lighting 36
2.2 Formats, color depth and compression 39
2.3 Resolution 43
2.4 In-camera sharpening 48
2.5 Color space 50
2.6 White balance 54
2.7 Storage media 56
2.8 Camera maintenance 58
3. Data optimization 63
3.1 EXIF 63
3.2 IPTC 64
3.3 Description - Naming conventions 65
4 Processing data 67
4.1 Importing and processing raw data 67
4.2 Tonal range levels and contrast adjustments 73
4.3 Adjusting resolution 78
4.4 Additional sharpening 79
5. Interchange format for the transfer of data 81
5.1 Recommended fle formats 81
5.2 Communication transfer 86
6. Application specifc considerations 89
6.1 Photojournalism 89
6.2 Portraiture photography 92
6.3 Advertising photography 94
7. License 97
Content
Image
Engineering
The overlapping communication of color according to ICC-Stan-
dard is todays foundation for color management in a media pro-
duction environment. Applied color management will bring
predictability and consequently security along into the color
workfow, from data input to conversion and output processing.

What is the fundamental challenge of color management within the
boundaries of physics:
- make the scan look like the original on the monitor
- make the print look like the original on the monitor
- make the print of a scan, without much examination on the monitor,
look very similar to the original
- generate verifcation for images to look similar on different machines
and monitors
Ten years after its introduction the vendor and platform independent
ICC-Standard has reached a high professional level and creates the
basis for an effcient color management system, in combination with a
progressive hard- and software development.
The entire communication of color in a modern production environ-
ment is going through a stage of modifcation, which yet already
takes place during photography due to the increased commitment to
digital photography. Whenever precise color reproduction is manda-
tory especially when shooting in a studio color management can
be applied in digital photography as such. When shooting outdoors
the vendors color transformations inside the camera will seize, for the
user color management starts once the fnal image will be displayed
on the monitor.
Working environment, general set-up
ICC Profles which characterize the color reproduction of a certain
device are being used in a controllable and safe process with image
data in order to transport color in a production process. The ICC speci-
fes device profles (ICC profles) and the mapping of the color engine
used, which realizes a consequent color management from photogra-
phy to print respectively displaying the images on the monitor.
The actual photography is very often the source for color data, hence
in the beginning there is the digitalization of a subject respectively
a scene or a scan. A modern working environment is based on the
1. Color management
1.1 Basics Color management
Image
Engineering
digital capture and reproduction of originals with a device or soft-
ware which supports color management and implements itself into
the workfow harmoniously.
Basic principles of color management / Conversion with ICC
profles
Communication in a standardized language is crucial for any kind of
management. Based on human color vision and comprehending all vi-
sual colors the device independent "CIE-L*a*b*" (see end of chapter)
color model is used as a standardized language, which describes an
equidistant color space based on human color reception. This is the
basic principle of color management.
Any input- and output device participating in the production process
(digital camera, scanner, monitor, printer, etc.) features an individual
characteristic way of color reproduction and describes a portion of
the colors which are visible for humans. This portion, which is being
referred to as "device color space", is visualized three dimensionally
as CIE-L*a*b* color space in the graph-
ic below. In order to transport color in
a standardized, comprehensive and
unique way through the production
process, ICC profles of every single de-
vice have to be created. The ICC profle
establishes the connection between the
according device dependent characteris-
tic (device specifc RGB or CMYK) and the
device independent characteristic CIE-
L*a*b* - in other words: The individual
color idioms of the individual input- and
output devices are translated into a mu-
tual language. Due to the fact that the
profle captures only a certain amount
of colors, an extra engine the so called
CMM (Color Management Module) is necessary for the calculation of
the secondary tonal values, one of each operation system, Apple and
Microsoft, already has one built in.
Due to the mutual language within the ICC profles it is now com-
monly defned, how certain devices see and reproduce certain color
values. Therefore single colors as well as complete images can be pro-
cessed and controlled as requested. For this purpose the ICC profles,
the CMMs and their implementation into the applications will merge
together and result in a color management workfow, in cooperation
Image
Engineering
with the special retouching-, editing- and proofng functionality.
Basic principles for data management / Workfow with ICC
profles, Interfaces
From the photographers view as the "data originator" data pro-
cessing and distribution in RGB is the better option, which one
while retouching can fall back to softproof functionality if desired
or requested in order to simulate the fnal output- or printed re-
sult. Needless to say the fnal output system needs to be familiar.

The input data is in relation to the complete workfow resembled as
digital photographs or scans, which will eventually get edited and
transferred either directly or afterwards
in company with graphic- respectively
text elements into a layout application
like Quark Xpress or Adobe InDesign in
order to be optimized and exported into
the fnal output process.

The photographer will initiate the
commission of data according to the
status of an image either to the post-
production or prepress interface.

The color- and process related optimi-
zation of data to the according output
process (in a normal case the separation
from RGB to CMYK) )can occur at differ-
ent sets of the workfow, while one generally needs to differentiate
between the process independent and process dependent approach.

The output optimization (separation) needs to take place at the end
when applying a medianeutral workfow and data management
which itself will generate perfect premises for the optimization to
further output processes. When choosing a process dependent ap-
proach the optimization respectively the commitment to specifed
output conditions will occur at an earlier stage in the workfow.

Both workfow options feature their advantages, with the
possibility for an optional hybrid approach at a later time.
To which extent the photographer will perform the output optimization
needs to be decided upon individually and to the standard of knowledge.
It shouldn't be regarded as the one primary exercise for the photographer.

The softproof-visualisation displays a pre-
view of the printed output "on-the-fy".
Image
Engineering
The integration of ICC color management to a working environ-
ment supplies these possibilities. Color management can make ac-
cess to generic or idealized color spaces or process descriptions
like ECI-RGB, ISOcoated, PDF/X,..... or depend upon adaptions
of special devices via individual ICC compatible procedures.

As one important component of workfow and data management, ap-
plied color management can safeguard adequate fexibility and security
to the according interfaces, much as the transfer of image material of
photographers to postproduction / prepress. It greatly matters to save the
current description of color in the shape of an ICC profle to every image.
(*) CIE = Commission Internationale de lEclairage. Defned in 1931 the

XYZ-color model and subsequently further standards, for example the
familiar CIE-standard color chart, known as the sole. This defnition is
based upon results of the latest research, from which a "standard ob-
server" was defned. This standard observer complies with 90% of the
color reception of humans. In 1976 the CIE-Lab color space was intro-
duced by the CIE, which in contrary to the "sole" had the advantage
of the calculative offset similarity.
Image
Engineering
1.2 Profling a digital camera
Advantages of camera profles
The advantage in the application of individually created camera pro-
fles lies within the compensation of "color-ametropia" (failure of vis-
ibility) of colors, which is inherent in any digital camera. The color-
ametropia differs from vendor to vendor as will discrepancies occur
between one vendors camera models. Through the use of camera
profles the photographer has more security, as a more precise color
reproduction from one camera as well as many different cameras,
instead of having to pass through long editing cycles in image editing
applications. Ultimately the use of camera profles reduces the hours
of work per image and therefore increases productivity.
Camera profle, more than a digital emulsion
Through the application of camera profles a digital camera is much
more qualifed for the precise reproduction of colors than the ana-
logue flm ever has been.
Just like the flm emulsions in analogue photography were designed
for different types of light it is essential for camera profles to be op-
timized for one type of light and should be applied with this only.
While in analogue photography the applied daylight- or tungsten
light flm in contrary to the used camera and optics displays the
most important parameter for color reproduction, the used camera
and processing technology and further workfow inside the camera
plays a major role, i.e. inside the raw data converter for color re-
production, which has to be incorporated when creating a profle.
Thus the camera profle describes the characteristic of a camera with
certain light as well as the adjacent camera- and data processing.
Furthermore not only daylight- or tungsten customizations can be re-
alized via camera profles, but as many "Illuminant-profles" as needed
for photographic work, i.e. a profle for HMI- or fuorescent tubes. The
profle can be applied to similar light types again and again just like
daylight or tungsten flm in analogue photography. For example a pro-
fle created for fash light in many cases can be applied on a daylight
location, when using the white balance functionality inside the cam-
era. This operation method will save the photographer a lot of effort
instead of having to create a new profle for every new composition.
Preferably the camera profle should only incorporate an exposure-
or contrast adaptation, if necessary due to insuffcient exposure or
gradation when taking the photo. The correct exposure and choice
of gradation will remain the photographers responsibility, just like
selective application of white balance. So mainly the camera profle is
supposed to correct the color reproduction of a camera.
Image
Engineering
Range of application for camera profles
This preliminary remark about the advantages of camera profles al-
ready indicates the range of application of such camera profles. The
application of camera profles plays a vital role whenever an exposure
depends on exact color reproduction. This could be art reproductions,
table-top and studio photography and whenever there is an increased
demand in color reproduction, e.g. for a corporate-, logo- or product
color. In photo journalism, where it is all about speed and changing
recording condition, in artistic photography as well as generally when
disposing a mood or feeling, color does represent a vital role, but the
precision of color plays an rather inferior role. In these cases the use of
individual camera profles is not categorically mandatory and due to
the customization of the workfow partly obstructive.
Creation of profles
Software and adequate test charts:
When creating camera profles special test charts are necessary besides
the profling software in order to capture the color characteristics of
the camera.
In principle the best test chart for this purpose is dependent on the
profling software. The most commonly distributed test charts for pro-
fling cameras are from Gretag Macbeth. For one thing the Digital
ColorCheckerSG and the ColorCheckerDC as well as the ColorChecker
with 24 color patches can be named here, whereas the ColorCheck-
erSG gathers the color characteristics of the camera best due to its
material properties. Some profling packages also use test charts pro-
duced on glossy photographic materials with additional greycards on
differing materials and can create good results. IT8 test charts used for
profling scanners are unqualifed for profling cameras. It is important
for the appropriate test charts to offer a good share of bright, semi
bright and dark colors, no optical brighteners as well as feature highly
saturated colors and therefore represent a
large color space. On the other hand the
colors are supposed to fll in the color space
consistently and a neutral grey balance (no
color cast under certain types of light) for
determination of the grey balance, which
will represent the grey characteristics of the
camera. A duller patch will prevent glossy
like refections of light more likely whereas
on the other hand the color characteristics
of the camera with highly saturated colors
can not be captured satisfactory. Hence a
semi matte material is a good compromise.
The Color Checker SG from Gretag
Macbeth
Compendium lor digilal pholography !0
Image
Engineering
Profling software should be able to operate according to most recent
ICC specifcations and use chromatic adaption after Bradford (math-
ematical function to adapt data to different light types) or newer
methods. The software needs to support different types of light and if
applicable add user-oriented types of light. Grey balance needs to be
able to operate in two different settings: choose a grey balance which
is calculated automatically from the software based on the data from
the test chart or not to affect the grey balance setting in the camera.

In addition the camera profling software needs to leave some options
to the user, in order to customise the profle according to individual
requirements. Just as different flm types were used in analogue pho-
tography in order to achieve perfect skin tones for example, the pro-
fling software should offer the opportunity to customise a camera
profle in relation to the sole purpose of its use.
Capturing the test chart and exposure
The most successful way to create a camera profle is to shoot the
test chart under repro illumination and preferably record with
maximum bit depth. This means to place the light source prefer-
ably two heads from left and right - in a 45 angle to the camera
and to the test chart, in order to illuminate the chart as plane and
evenly as possible. The ideal condition for capturing the test chart is
a darkened room with no disturbing refections on the chart as well
as light source with constant temperature, illumination of light and
equal brightness. Likewise the test chart can be captured outside
the studio as well if disturbing refections and mirroring can be pre-
vented. Refections can be minimized by underlaying or hanging a
matte black or grey piece of cardboard or cloth of adequate size.

The right exposure for the test chart is essential for the success of a
premium digital camera profle. If the exposure of the test chart is
darker than that of the images you will be taking, the profle will
make the images look much brighter. In reverse a brighter expo-
sure of the test chart will end in a darker reproduction of your im-
ages. The correct exposure should best be determined by a row of
exposures. A reference value for the correct exposure when using
the Gretag Macbeth test charts can be checked upon by the white
test chart patches. The patch should keep lower RGB values than
the white areas of your image with just as little detail as possible.
When using premium studio cameras the right exposure of the test
chart should result in RGB values from 235-245 in the white patches.
For SLR and rangefnder cameras a good starting point for the right
exposure of the test chart is measuring a greycard (e.g. the Kodak-
!! Compendium lor digilal pholography
Image
Engineering
greycard) and transferring the exposure values to the test chart shot.

Since color-ametropia shows different effects with different lighting
conditions, it would be self-evident to place the test chart directly into
the scene to be photographed and create the profle from there. Prac-
tical experience shows little success for this approach. The illumination
of the scene very often leads to e.g. black patches of the test chart to
be too light as a result of direct refections of a light source and there-
fore will create errors in the profle. The same is obviously true for
the patches that refect directly into the direction of the camera. Add
occurring falsifcation of colors to this, due to refections of colored
objects in the scene. For example, when shooting a tabletop setup on
a colored underground, the refections of the underground will inter-
fere and lead to falsifed colors in the test chart and to bad profles.
Likewise with refections of metal objects or white walls which can
infuence the capturing of the test chart in a negative way.
White balance
A correct white balance is one of the most essential requirements for a
good profle as well as the subsequent application of the profle. The
white balance can be comprehended as a calibration of the camera to
the according lighting conditions. It must be pointed out that some
SLRs and Rangefnder cameras will accomplish the white balance by a
white area. If using an improper material like a piece of paper for this
it may lead to drastic errors in the profle due to optical brighteners in
many papers. An alternative for the correct white balance with a larg-
er, neutral area is the use of the Gretag Macbeth White balance Card.
The predominant portion of cameras will acquire white balance of
a neutral grey area. There are full-format greycards available on the
market for this. A Kodak greycard with its 18% refection is very ap-
propriate for exposure metering, but not for white balance, since it is
not suffciently neutral.

Depending on the camera the greycard needs to be captured full-frame.
The grey patches of the camera chart can be utilized for white balance
as well for more premium cameras. Double check the white balance in
the image via the pipette of your image editing application or the raw
data converter. A successful white balance will favour an interpretation
of the camera of neutral grey areas in the image with equal RGB values.

If the white balance, the exposure and the illumination of the test
chart have been performed according to these instructions, then you
have made a foundation for a good camera profle. Please consider
during the application of the camera profle that the profle will oper-
Image
Engineering
ate reliably just as long as the basic settings in the camera respectively
in the further editing workfow are maintained between the image
capturing and the capturing of the test chart. If e.g. switching the
raw data converter, the profle created with a different converter can
not be applied anymore. Furthermore, always keep in mind to double
check white balance as an essential quality feature during image cap-
ture and modify it if necessary, then you are on the safe side, to use
the one-time created camera profles for a lot of shots.
!3 Compendium lor digilal pholography
Image
Engineering
In image editing applications the effect of certain tools on image data
is infuenced by the color mode used one reason to look into color
modes and color spaces a little closer.
It is an interesting experiment to perform the same adjustment of im-
age contrast once in RGB, CMYK and Lab. When increasing the con-
trast of the image a little by lowering the shadows and raising the
highlights you will be able to experience dedicated image material
presumed the following effect:
RGB color mode: saturation of colors is increased.
CMYK color mode: saturation of colors is increased but color shifting
occurs and the shadows will fade out.
LAB color mode: no shifting of colors but increasing contrast due to the
separation of the brightness information from the color information.
Besides the investigated effects other factors need to be taken into
consideration for the choice of the appropriate color mode. The con-
version of data to CMYK (= separation) requires accurate knowledge
about the fnal print process and the necessary color removal. The use
off LAB in turn will lead to loss of details with 8-bit color depth in criti-
cal areas of the image; Furthermore some tools in e.g. Photoshop are
not even available in LAB color mode (e.g. selective color correction).
It turns out that RGB color mode is the most reasonable one for edit-
ing digital images.
From color mode to color space ...
It seems even more confusing since there is a multitude of color spaces
available for the RGB color mode. A color space is a certain RGB de-
scribed by means of an ICC profle. At the beginning of the job the user
faces the challenge of choosing the correct color space for the captur-
ing, editing and archiving of their images. The color space chosen for
the editing of the image (image retouching, image enhancements,
unsharpmasking...) is being referred to as a working space. (Attention:
Once again this is not the RGB color mode but a color space character-
ized via an ICC profle, e.g. eciRGB or AdobeRGB!) Prior to the decision
for a certain working space, some issues need to be considered.
Working spaces are described by the following parameters:
- the color values of the primary colors red, green and blue
- the gamma value
- the whitepoint
Primary colors
The defnition of the primary colors affects the size of the color space
in the frst place. One needs to pay attention for a big enough result-
ing color space for the desired output process, since during the cre-
1.3 Working spaces
Image
Engineering
ation process or during conversion to the working space those colors
will be clipped already, which could have been useful in a later high
quality premium output process (e.g. for Gravure or Offset printing
processes). In order visualize the different color spaces graphically the
2- or 3-dimensional illustration of the "ColorSync service application"
under Mac OSX can be used. For all others a standard browser with
VRML-plugin can be used for a comparison of the size of two color
spaces via the website http://www.ICCview.de/.
Gamma
The gamma value defnes how the gradient of brightness is based on
the profle, or how many values at a time will describe the highlights
and shadows. The lower the gamma, the more values are used to de-
scribe the differentiation of the highlights and if higher the more are
used to describe the areas in the shadows. In practical use a gamma
of either 1,8 or 2,2 can be found, whereas we fnd a slight preference
for a gamma of 1,8 in color spaces coming from the area of prepress
and a 2,2 in the display areas. Recently there are also such color spaces,
which replace the gamma with the brightness distribution of the LAB
color space which is equivalent to the reception of the human eye (L*,
pronounced L-star).
White point
The "color" of white in a profle is described by the Whitepoint. In this
case white is an area refecting light over the range of visual colors.
This color is primarily dependent on the illuminating light source. Two
"normative" light sources have been established along with the ac-
cording Whitepoints: D65 and D50. These describe the coloration of a
reference object at a temperature of 6500K and 5000K. Since the illu-
mination of D50 has been established in prepress and print it is useful
to apply a working space with D50 and to calibrate the monitor to it,
if you are working in this area. (Also see chapter 1.4 "Viewing condi-
tions" and 1.5 "monitor calibration".)
General considerations when choosing a working space
Those color spaces describing an input- or output device, are of
limited use as a working space. Because of its "nonlinearity".
This means that a color with equal RGB values (e.g. R=G=B=127)
will NOT result in a neutral grey and therefore e.g. optimiza-
tions in contrast will occur accompanied by a distinct color cast.
Thus those working spaces are especially favored which feature nor-
malized RGB color spaces and will generate a visually neutral grey.
Image
Engineering
Appropriate working spaces
A brief while ago the European Color Initiative (ECI) aimed at the de-
velopment of an RGB color space, which will cover the current estab-
lished printing color spaces as good as possible, without becoming too
"big". Since the Apple operating system took a commanding position,
a gamma of 1.8 was chosen. As a Whitepoint D50 was specifed and
the primary colors were ajared to the NTSC specifcations. The ECI color
space is very well suited for editing, delivering and archival of image
data, since the color space covers all current printing color spaces fairly
well. ECI-RGB can be downloaded free of charge from the website of
the ECI (www.eci.org). If you are working in an environment though
that does not support color management please bear in mind not to
use any image data in eciRGB that has not been converted prior to
application. These are e.g. webapplications and Offce environments.
For this purpose a conversion to sRGB beforehand is recommended.
Another meaningful working space is AdobeRGB, since it features
the same gamma value as sRGB and some applications and cam-
eras as well use it as a default setting. Especially with software
products that do not attach color profles we need to pay good
attention to the compliance of camera software working space
and image editing working space. Additionally unnecessary col-
or space changes should be prevented, since the loss of quality
due to multiple conversions can not be eliminated to full degree.
(Note: Older Nikon digital-SLRs apply an inferior color reproduc-
tion in AdobeRGB than in sRGB. In order to ensure a trouble-
free workfow one should always conduct a number of color
tests prior to switching the working space of the digital camera.)
There is an interesting approach to this from PhotogamutRGB (http://
www.photogamut.org/) The idea was trying to cover all current pho-
tographic output- and printing processes and to eliminate all colors
not contained from this working space. On top of everything a har-
monization with sRGB was aspired in order to reduce color falsifca-
tions due to the transfer of image data into a workfow without color
management. Please note the application of a so called LUT-profle in
order to characterize this color space, which is over 150 kb in size and
therefore increases the fle size to this amount. It has not spread in
practice so far, but its defnitely worth keeping an eye on its develop-
ment.
Special color spaces
We have not dealt with the extra big color spaces like CIE-RGB
or WideGamutRGB up to this point. Those cover a visible range
as big as possible but should be applied in special cases only. It
must be pointed out that the quality of a future color separa-
Image
Engineering
tion will be higher if the size of the working space and the desig-
nated target color space will not differ signifcantly. Additional-
ly these color spaces can not be displayed any more on a monitor.
Note: It is not possible to display working spaces like eciRGB, Ado-
beRGB, PhotogamutRGB on a standard monitor at the moment. This
means it is not possible to visualize the differentiations in highly satu-
rated colors. Some special software applications for monitor calibra-
tion offer a workaround (which can partly eliminate this drawback) as
well as some special monitors. These monitors have only been avail-
able for a short while and distinguish themselves with a large color
space besides a rather high price of cost beyond Euro 4.000,-.
Color spaces for Internet- and Consumer applications
When producing and editing data for in-
ternet applications a very special approach
has proven itself, which takes the peculiari-
ties of Internet applications and "Consum-
er monitors" and printers into account.
Normally Web- and Offce applications
will not support color management. In
order to be able to communicate re-
sembling colors to some degree, the IEC
with the cooperation of some rather big
manufacturers (HP, Kodak, Microsoft)
specifed the sRGB color space not so long
ago. Normally monitors and consumer
desktop printers are appointed to repro-
duce the sRGB color space close enough.
If images are placed in Offce documents
or in a web browser, a true-to-color re-
production can be achieved best, if the
data has been characterized with sRGB
as well. So there is the following recom-
mendation for Offce- and internet appli-
cations: Choose sRGB as output working
space. This color is best suited for most
applications, which do not support color management, for example
Offce software or generally the Windows operating system. In addi-
tion many photolabs that offer the output of digital data on paper
prints will not take attached ICC-profles into account but simply as-
sume that the data was processed to sRGB. Many digital cameras will
produce sRGB image data by factory default setting. Some more ad-
vanced cameras will sometimes offer a different color space. It needs
to be pointed out that the use of sRGB as a working space is not rec-
Comparison of color spaces:
Yellow indicates Photogamut RGB, red ECI-
RGB, grey and almost as big as ECI indicates
Adobe RGB and the smaller color space within
is sRGB.
Image
Engineering
ommended in a professional like environment, since sRGB describes a
relatively small color space, leaving out some colors that can very well
print in e.g. Offset printing process (particularly in the Cyan range, but
in yellow as well).
In Conclusion
There is no ideal working space for everything, but there is an appro-
priate one for every intended application. In the premium prepress
area eciRGB is a good alternative, AdobeRGB is a well spread world-
wide standard, which can already be provided by some input devices
like scanners or digital cameras. But there are already some new inno-
vative solutions emerging on the horizon, which will eliminate some
of the disadvantages of todays well spread working spaces. Some ad-
vice at last:
1.) Exclusively use calibrated and profled monitors for displaying
your images at all times and a software which supports color man-
agement for displaying.
2.) Stay well away from editing in CMYK unless the client resists on it.
3.) Always attach the ICC profle used when saving the image data.
4.) Pay attention when choosing your designated output space to be
covered fully by your working space.
Image
Engineering
It is due to the fact that digital data can not be displayed without the
aid of such devices as a monitor or printer, which makes calibrating
and profling the device used essential since data will not be displayed
correctly otherwise.
Just as important as the calibration and profling of these devices
are the viewing conditions under which the images will be viewed.
It is rather clear that these conditions can not be ideal at all the times
for example for situations like in the area of photojournalism or when
shooting "on location". It gets rather important for the photographer
to have the kind of experience at hand in these situations in order to
estimate the kind of effects these conditions will have on the image
displayed. This guide can not provide this kind of experience but it
can highlight the conditions, regarded as ideal when viewing images.
Those of you wanting to take this one step further can have a look
into ISO 3664 (Viewing conditions Graphic technology and photog-
raphy).
Walls and luminance
The walls should be painted either white or grey. A colorful paint-
work or even colored foors need to be avoided. The luminance of
the surroundings around the monitor should be according from
60% to 10% of an white area on the monitor. Illumination needs
to be as constant and even as possible. If daylight is used for il-
lumination there needs to be some sort of protection from direct
sunlight. Aluminum blinds are suited best for this purpose if their
lamellas are rotatable and arranged horizontally. The lamellas are
color neutral and if using the appropriate angle the light can be di-
rected to the ceiling and the amount of light can be adjusted. This
will provide the ideal illumination without disturbing refections.

Tip: Point a lightmeter with attached calotte to a white area on
your monitor and measure illumination. Than put the lightmeter
against a white wall and measure the amount of light falling onto
the wall with the attached calotte (not the refecting light). The
monitor needs to be two f-stops (range of 1-3 f-stops is o.k.) brighter.

Try to avoid scattered light and direct refections on the monitor. This will
work best by attaching the lights (normlight fuorescent tubes) paral-
lel to the window and equipped with glare shields. The monitor needs
to be aligned parallel to the window area in order to prevent refec-
tions and scattered light and should be equipped with a monitor hood.
1.4 Viewing conditions
Image
Engineering
Tip:
If there was no such hood included with your monitor, you can custom
build one yourself (make them 35 cm wide) with a couple of black ex-
panded plastic slabs, available from any specialized vendor for graphi-
cal supplies, and attach them to the monitor with some velcro. The
cost adds up to something like Euro 3,- and they are just as good as the
fancy accessory from the vendor.
Quality of light
Daylight is nice for illumination but it changes its color (spectral com-
position) in relation to the weather situation and time of the day.
Halogen lamps, light bulbs
and regular fuorescent
tubes are not appropriate
enough for the illumina-
tion of an image editing
workstation. The illumina-
tion of choice are fuores-
cent tubes for standardized
lighting, thus fuorescent
tubes with a spectral com-
position, which gets very
close to natural daylight. It
doesnt have to be the ex-
pensive special tube from
the manufacturers of view-
ing boxes for proofng purposes. The tubes from Osram or Philips
will do the job very well and dont cost a lot more than a regular
tube in the electronics department store. Look after the classifca-
tion from Osram "Lumilux de Luxe xxWatt/950" and from Philips
"TLD de Luxe xxWatt 950". Thereby the most important criteria is
the 950 whereas the 9 stands for the highest quality of light and the
50 stands for 5000K. The power varies with the length of the tube.
With the release of this guide Philips will bring a special TLD tube with
the classifcation TL-D 90 Graphica Pro, which can be recommended.

Those who need it precise can use a standardized color viewing box,
available e.g. from Just Normlicht, Kaiser Fototechnik or Mega Stahl.
One and the same color looks quite different under
different viewing conditions. This is why the viewing
conditions are so important.
Image
Engineering
As a user we must not forget that we can not actually see digital imag-
es with their "zeros and ones". We will always need an utility for the
visualization, like e.g. a monitor or a printer. We will never achieve a
correct presentation on the monitor if the utility has been misadjust-
ed (like too dark or too contrasted). But what actually means a correct
presentation in this context?
As a start we need to conceive the fact, that a correct presentation of
digital images is only possible when using image editing applications
that support color management according to ICC specifcations. As
described in chapter 1.3, the image will be edited within a working
space in this application. The description of this working space has
to be included with the image in the shape of an ICC profle. The im-
age editing application knows the color of each pixel from the digital
value of each pixel and the according ICC profle (discrepancies from
color management modules and rendering intents are left ignored at
this time). For the pixel to be presented correctly on the monitor, the
image editing application needs to know which digital value to send
to the graphics card, in order to create the same color. This occurs
likewise via an ICC profle for the monitor. The following will describe
the process of creating such a profle.
Calibration of the monitor
First of all the monitor needs to be calibrated. What does that mean?
A monitor that has been set e.g. too dark or too dull will never be
able to display an image correctly. This means the monitor, in order to
display as many colors as possible, needs to have a color space as big as
possible and needs to be adjusted perfectly from the hardware side.
Part of these optimum settings are besides the geometry, brightness,
contrast and whitepoint in particular.
While brightness and contrast can be checked with test images avail-
able in tools like the DQ-tool from the german association of photo
industries, on the ECI website or in software from BasICColor or X-Rite,
a spectralphotometer or at least a 3- or 4-range measuring device is
necessary to analyze the whitepoint.
Also the presetting of the brightness characteristics, so far this has
been called the gamma of the monitor, is part of the calibration pro-
cess. This gamma is based on the mathematical description between
the interrelation of the input signal to the monitor in the shape of
voltage and the emitted amount of light (see link to Charles Poyton).
Since this gamma will be managed from the graphics card or the mon-
itor electronics of modern computers/monitor, the usual principles are
not valid anymore for the control of todays LC-displays, the calibration
1.5 Visualisation on the monitor
2! Compendium lor digilal pholography
Image
Engineering
with the help of a gamma is currently replaced by a new optimized de-
scription. The calibration towards a linear L* is being discussed in this
context. With this the L-axis (brightness-axis) of the Lab color space
is meant, which describes the lightness reception of the human eye.
Therefore the monitor should be set up as ideal as possible for the
human eye with a calibration according to
L*. If the software for monitor calibration
does not offer a way to calibrate according
to L*, then a gamma should be chosen, that
comes closest to the human perception and
that is a gamma of 2,2. This setting can be
used on the Macintosh as well as on the
PC. The historical reasons for setting the
Macintosh Gamma to 1,8 came from the
prepress area and presently have no more
relevance.
The choice of the whitepoint is another
controversial issue when calibrating a monitor. The ideal whitepoint is
the one that comes as close as possible to the regular viewing condi-
tions for images. A study from Kodak proves the spectral composition
indoors tends towards a color temperature of 5000K. In addition there
are special sources of light with a spectral composition based on the
color reception of 5000K respectively a standardized light source with
the classifcation for D50. That means to set a whitepoint of D50 on the
monitor. There is an exemption for CRT-monitors and especially older
monitors which due to their light density contrast range will produce
a very yellowish presentation when choosing D50 as a whitepoint. If
you have such a monitor, you better choose 6500K as color tempera-
ture respectively the whitepoint of D65. An image, displayed on such
a monitor will come across a little bit cooler in direct comparison with
the original under D50 standardized light, but matches rather well to
the original relative to the white areas.
The ideal settings for the monitor calibration are according to this:
check brightness and contrast with a test image or with the aid of a
measuring device.
Choose the luminance characteristics if possible according to L*, oth-
erwise choose gamma 2,2.
Choose 5000K as a whitepoint and 6500K only in the mentioned ex-
emptions.
Creating an ICC profle
Subsequently to the calibration an ICC profle has to be created for the
monitor. This can be accomplished with the aid of a monitor measur-
ing device and the according software. In the process of the profle
Image
Engineering
creation different colors will be displayed on the monitor and then
measured by the device attached to it. A profle will be created from
the data collected and needs to be set as monitor profle by default
in the according operating system. Applications like Photoshop will
automatically apply the profle in order to display the images correctly
on the monitor. Ever since Photoshop this can fortunately not get de-
activated anymore and the user will always get the true-to-color pre-
sentation, provided the color settings in Photoshop are correct (see
chapter 1.7).
Take caution with Adobe Gamma and similar tools
In the past the premium quality measuring devices were rather expen-
sive which is why there were some attempts to accomplish the calibra-
tion of the monitor with the help of simple auxiliary tools. Therefore
Adobe offered a small application called "Adobe Gamma" and Apple
within their monitor settings. These applications were operating by
the "rule of thumb" principle and a profes-
sional at the controllers could achieve a half-
way reasonable result. If an untrained user
operates the controllers catastrophic results
were rather often the results. The use of a
monitor measuring device can prevent that.
Currently a good monitor measuring device
will cost from 250,- plus and it really is a
must for everybody, who deals profession-
ally with images. Without such a measuring
device, displaying images on a monitor is
like fying blind.
Softproof
With the aid of a monitor profle the user will always receive a cor-
rect presentation of the image on the monitor (naturally only within
the physical boundaries), regardless of the color space the image is in.
Thus it can be available in different RGB working spaces, in a device
color space o even as CMYK in an printing process color space. In or-
der for the profle to be actually used by applications like Photoshop,
it has to be set as system profle for the monitor. This can be achieved
on the Mac under "System preferences/Displays/Color and on the PC
under "Control panel/Display/Settings/Advanced/Color Management.
Most monitor measuring devices will set the profle automatically as
Systemprofle after calibration and profling. A little supervision whe-
ther the newly created profle really was set in the System settings is
Image
Engineering
ok anyway.
Since the monitor respectively the image editing application doesnt
care about in which color space the image is in, it can additionally
simulate the output on a different device if there is a description of it
in the shape of an profle available. This simulation can be chosen in
photoshop under the menu for Ansicht/Proof einrichten/Eigene (see
also chapter 1.7). In this case the images will get converted to the
chosen simulation color space in the background and only then get
displayed on the monitor. Thereby the displayed image will be limited
to the range, which can be produced by the output device.
This pictue shows
the original RGB
fle on the left
and the softproof
presentation for
the offset output
process on the
right hand side
Image
Engineering
In the range of digital photography for consumers we can not auto-
matically assume the user to be familiar with things like color manage-
ment. Thus all devices in that range are set in the way for all recorded
images to be displayed fairly neat on an average monitor. Likewise
the printer drivers of all common inkjet printers are tuned in the way
all images printed will look halfway like the image was seen on the
monitor. Since the sRGB working space comes very close to an average
monitor there are usually very little problems in consumer workfows
but not necessarily the best results possible.
The feld of professional photography applies different working spac-
es (see chapter 1.3) and there is a demand for a higher precision. The
image displayed on the monitor is supposed to match the printout as
close as possible. This can only be achieved within the physical bound-
aries of the according output device / printer. But the wishes can come
true within these boundaries.
Profling
It is essential to have a profle for the according output device, which
is characteristic for the device-material-combination used (printer, pa-
per, inks, driver). In the forefront it is important to calibrate the output
device. It is supposed to deliver optimum results and constant output
over the time. This is not working for some Color copiers since the out-
put result is dependant on air temperature and humidity. But modern
ink jet printers deliver a constant result, as long as the air humidity
does not exceed 70%. For the creation of a printer profle a color chart
will get printed depending on the according driver software which
will be measured subsequently with a spectralphotometer. If you are
afraid of the investment for such a device you will fnd a professional
service provider in every big city, who will create the profles for you.
Possibly Google can be of help fnding one --> search for: ICC remote
"profling service" to get a list. Just send the printed charts to one of
the service providers and you will receive the created profle back. This
needs to be copied into the profle directory of the according operat-
ing system in order for applications like Photoshop to recognize the
profle. TIP: under Windows XP just right-click the profle and choose
"Install profle".
Preview of the print
Since most printers are limited in color reproduction due to paper and
ink combinations used, it is good to display the fnal result of the print
in advance. Photoshop can realize this with the help of the profle.
The exact procedure will be described in the next chapter.
1.6 Visualisation in print
Image
Engineering
Print with Preview
In order to print correctly the printer profle needs to be chosen in the
"Print with Preview" menu. Thus will convert the image before send-
ing it to the printer driver in the way that it can be printed correctly.
This printing alternative is called "Full Gamut Print" since it takes full
advantage of the printers color space. At this point a so called Proof
can be activated as well in order to simulate different output process-
es on the according printer like e.g. offset printing for that purpose.
To exercise control over whether the simulation of the printing pro-
cess was successful, there is the Mediastandard print for offset print-
ing which is based on the mediawedge published by the FOGRA and
UGRA. It is about color patches that will be printed and measured with
a spectrophotometer. Each and every one of the color patches respec-
tively certain groups of them have to comply with certain tolerances,
so that the print can be regarded as compliant to the offset print.
A similar process for the control over RGB image data is under stan-
dardization by the DIN.
Print with Preview Dialogue above. An output profle can be choosen. Below
the Mediawedge (more information from www.fogra.org)
Image
Engineering
Printing with a RIP
From applications like Photoshop images can be printed trouble-free
with simple profles. A RIP (Raster Image Processor) is essenttial if a
layout with graphics, images and text is printed from applications like
Quark Xpress or inDesign. Such a RIP can be avoided by creating a PDF
fle and printing it directly from applications like Adobe Acrobat.
In the professional area this workaround can be too time-consuming
most of the times and the image setters for the printig flms and pla-
tes still require the RIPs. They do have a couple more advantages. For
example regular printjobs can be collected and printed space-saving
or big posters can be split into a lot of smaller prints.
When planning to purchase such a RIP pay good attention whether
it supports ICC color management. The profling of such a RIP is a bit
more complicated since they can adress the printer colors directly. This
will initially require e.g. the limitation for color-application and a line-
arisation. If you are planning to build a profle for your printer-paper-
RIP combination you might want to consider to consult an expert.
Image
Engineering
Compendium for digital photography
Image
Engineering
Postcard:100 x 150 mm
US-executive: 184 x 267 mm
DIN A4: 210 x 297 mm
US-letter: 216 x 280 mm
US-legal: 216 x 356 mm
SUPER A4: 241 x 340 mm
US-tabloid: 280 x 432 mm
DIN A3: 297 x 420 mm
A3 F: 305 x 457 mm
SRA 3: 320 x 450 mm
SUPER A3: 328 x 453 mm
DIN A3+: 330 x 483 mm
A3++: 329 x 558 mm
DIN A2: 420 x 594 mm
DIN A2+: 480 x 628 mm
DIN A1: 594 x 841 mm
DIN A1+: 625 x 914 mm
DIN A0: 841 x 1189 mm
DIN A0++: 914 x 1250 mm
Tecco Format Guide
www.tecco.de
Image
Engineering
At frst we fnd that color settings in DTP applications via pro-
gram settings are being discussed quite controversial in some ar-
eas and even among experts. Certain "Sets" have yet been prov-
en in practice in such ways to reduce the probability of "color
accidents" - thus can not totally prevent these. Furthermore in the
settings only such profles are being used, which allow for stan-
dardised production processes, and this should not be under-
estimated are available free of cost to the user (www.eci.org).
Image editing can not be imagined today without Adobe Photoshop
and thus shall be used for the choice of settings as an example. These
can be transferred to other applications supporting color management.
With the completion of this guideline Adobe Photoshop is avail-
able in version "CS" or "8.01" for Windows (2000,XP) and MacOSX
(10.3) as a component of the "Creative Suite". Adobe has attached
great importance to the compatibility of both platforms during
the development. This was accomplished quite well both on the
frst view (user interfaces, menus, palettes) as well as "under the
hood". Particularly the handling of colors and the important as-
sociated profles will lead to roughly the same results on Apple
as well as on windows-based systems. Differences can be found
mainly for the locations of settings we will cater to this later.
The screen captures used in the explanation were taken under OSX .
The illustration doesnt differ fundamentally from windows systems.
Color settings
The color settings occupy a central importance for the presenta-
tion and internal conversion of colors, and will be opened via the
menu item "Photoshop"-> "Color settings". It is important to acti-
vate the expert mode frst. The dialogue is divided into fve differ-
ent zones: settings for the working spaces, association of the pro-
fles (color management guidelines), the conversion rules for the
CMM and eventually the extended settings and the descriptions.
The ECI gives out the following recommendations for a working
space: As RGB working space eciRGB should be used. This working
space is very well suited for the elaboration of image material and
has achieved great acceptance in central Europe on a broad basis.
Furthermore eciRGB covers all colors that can be reproduced with all
current output / printing processes. Under special circumstances e.g.
the exclusive use of digital SLR cameras as a source for images this
recommendation can be diverged from. ( see 1.3). Choosing a CMYK
working space depends on the subsequent assignment. The ISOcoated
proved itself in most cases. This color space describes the standardised
1.7 Settings in Photoshop
Image
Engineering
printing process according to ISO-process norm 12647-2 for coated
and noncoated papers in a sheet offset printing process. ISOcoated
is also recommended for the presentation of greyscales and spot col-
ors. (Note: This can be achieved by selecting the ISOcoated.ICC pro-
fle from the popup-menu next to the
according working space and choosing
"load greyscale settings" or "load spot
colors"). If the optimization of data for a
printing process is neither necessary nor
wanted, than "Grey Gamma 2.2" needs
to be chosen as greyscale working space.
For the color management guidelines the
setting "keep embedded profle" should
be maintained . "Profle errors" and
"missing profles" defne the interaction
with missing profles or profles that do
not comply with the chosen working
space profle. At this point, there is the
recommendation to select all options.
This will trigger photoshop to open a
dialogue whenever an image does not
contain a profle or a profle does not comply with the chosen working
space and also the user will get the opportunity to convert into the work-
ing space or to assign an alternative profle when opening the image.
For the conversion options the setting "Adobe (ACE)" has proven itself
due to the uniform handling on different platforms as well as the "percep-
tive" setting. Blackpointcompensation as well as Dithering can be acti-
vated as well. In the extended settings no options are necessary to choose.
Sure enough all these settings appear complicated enough, but the
ECI has come up with a solution to this problem and publishes in-
stallation applications on their homepage, which offers ready-to-run
color settings. These can be activated via the topmost popup-menu
"Settings" and already contain the ICC profles for the working spaces
thus wont need to be installed separatly. In conclusion the standard
installation presets (e.g. Standard settings for europe" should not be
chosen.
Opening image data
When opening image data Photoshops be-
haviour will vary after setting the according
working spaces. The follwing scenarios may
occur:
1.) No profles embedded: If the image
data does not contain a profle, Photo-
Image
Engineering
shop will show the following dialogue while opening: The user ob-
tains the options to either open the image without a profle (in this
case the according working space will
be used in order to visualize the RGB- or
CMYK-values), assign a working space
or assign an alternative profle and op-
tionally convert to the working space.
2.) Profle embedded that does not comply
with the working spaces: The embedded
profle does not comply with the according
working space (RGB or CMYK) in Photoshop
while opening an image, Photoshop will show the adjacent dialogue.
The user has the option to either keep the embedded profle, to convert
the image data to the working space directly or to discard the profle.
3.) Embedded profle matches with working space: If the the embedded
profles in the image data match the working spaces (RGB or CMYK),
Photoshop will accept these image data without further notice, pro-
vided that no RAW-data is involved ( see also Chpt. 4.1 and 5.1 Raw
data; the behaviour of Photoshop will be described further down).

With this multitude of options, it is essential to fnd the appropriate
strategy fur the further processing of images. In principle the are op-
tions, to assign a profle (which means that the pixel values in the
image will NOT get changed but the the colors displayed will vary
greatly) or to convert (into the working space, in this case the pixel
values will be CONVERTED, thus CHANGED the display on the moni-
tor probably changes marginally). A generally accepted declaration
about which strategyis better can not be made at this point. It seems
to make sense though, when opening an mage to merely ASSIGN a
profle, which describes the image accordingly (keep embedded pro-
fle). This assigned profle can still be changed at a later time when
CONVERTING this is nopt possible anymore. This approach should be
preferred also since there is no visual feedback (preview of the chang-
es) possible on the monitor. On the other hand if a different profle
will be assigned via the menu item "Image -> Mode -> assign profle"
the changes will become visible when activating "Preview".The same
does apply to the conversion of image data via the menu item "Image
-> mode -> covert to profle".
Opening rawfles (raw data)
0.The opening of rawfles (see also Chapter 4.1 and 5.1) takes an
exceptional position. The so called "Camera RAW plugin" comes "in-
the-box" with Photoshop 8, which enables opening rawfles directly.
The following strategy has proven the test of time:
Image
Engineering
1.) determine white balance / defne neutral grey
While opening photoshop will use the white balance determined by
the camera. This can be noticed in the right menu area under the
"view" tab throught the white
balance preset "As camera". If
the white balance is not pleasing
enough (neutral areas e.g. show
an explicit color cast) a change
can be applied in the frst place
throug either a global modif-
cationby selecting "Auto" (or
the type of light available while
capturing). If this approach will
not provide satisfying results,
changing the color temperatur
(lower values --> image will turn
bluish, higher values --> image
will turn yellowish) and the tint (negative values --> image will trun
greenish, positive values --> image will turn magenta) can obtain a
remarkable optimisation.
A further option is to adjust to a neutralgrey area in the image itself.
This will be acieved by choosing the grey balancetools in the win-
dow on the left and selecting (click) the the area of the image to be
matched. Therupon Photoshop the colortemperature settings and
the tint until this tint is defned as neutral (Comment: If "AdobeRGB"
is selected, neutrl grey areas of the image will always hold values
of the color components R=G=B. If the mouse pointer is within the
image the values will be shown on the right side unter the image
window).
2.) Defne Exposure
After adjusting the whitepoint respectively the defnition of neutral
grey it is recommende to defne the exposure. The highlights in the
image will be infuenced with the Ruler "Exposure", the shadows in
the image with the shadows ruler. When holding down the "Alt"-key
while dragging the ruler, Photoshop will highlight the parts in the
image that are assigned as white ("Exposure" ruler)and hich parts of
the image are assigned as "black" (shadows ruler).
Comment: These visible parts of the image shown
in these demonstrations will hold no picture details
whatsoever, therefore they are "fat" or "clipped".
3.) Adjust brightness, contrast and saturation
Choose the whitepoint either from an entry in the list, by
moving the slider or via clicking the pipette within the
image.
Image
Engineering
The rulers for "brightness", "contrast" and "satuation" will accom-
plish a further adjustment of the image itself. The functionality of
these rulers is rather familiar in the fled of color corrections and
does not require further explanation.
4.) Apply sharpening
Sharpness can beapllied to the images optionally even while import-
ing the images . The sharpness settings can be found in the right
menu area underneath the "details" tab. Please note in order to
judgesharpness appropriatly a 100% zoom isessential. Therefore it
is rcommended to change the zoom to100% before changing the
sharpening parameters.
5.) Set parameters for the conversion
Before actually converting the image a few parameters need to be
defned. "AdobeRGB" can be recommended as color space at least
as long as Adobe Photoshop does not offer atlernatives for Working
spaces like ECI-RGB" or "PhotogamutRGB" in the Raw-plugin.
If the image is already according the the users anticipation by now
and no further optimisation of tonal range and colors is necessary in
Photoshop, it can be openend with 8.bit depth. In all other cases the
image should be opened with 16-bit color depth.
The menu item "Size" offers depending on the typeof camera used
different options to either enlarge or reduce the image size. The
quality achieved is clearly superior to the regular scaling of images in
Photoshop. Now if its clear while importing the data to use it for a
large-format-poster, then the scaling should be apllied in the RAW-
plugin.
300 pixel per inch have proven to be appropriate resolution since it is
very close to the theoretical resolution limit of the human eye. Once
the settings have been aproved with "ok" Photoshop will calculate
the fnal image from the camera data. This can take a couple of
seconds according to the amount of data and theperformance of the
PC or Mac.
Visualisation
Photoshop CS makes a couple of technically matured tools available
for the visualisation of images.
An appropriatly calibrated and profled monitor is a requirement in
any case, which can display a color space a big as possible (see alo in
chapter 1.5 Visualisation the monitor).
When visualizing colors we need to consider the fact, that the moni-
tors and displays used today are NOT able to display all colors of an
RGB-fle (e.g. in ECI-RGB or even in sRGB) correctly. This is also true for
the printing colors used in an offset printing process even though
the offset printing color space according to ISO 12647-2 for uncoated
Image
Engineering
paper is signifcantly smaller than the RGB color spaces mentioned
above, a highend monitor is necessary in order to display all colors
available in print on the monitor.
If a visualisation of animage in a certain colo space is conducted, only
those colors will be shown which the monitor is able to display.
But we have to fnd in relativation that the majority of image ma-
terial used will not exploit the color space of neither print nor RGB
color space. Thus in most cases the visualisation of colors on the
monitor will give us a pretty good impression of te fnal results.
If animage is availble in a RGB color space, the colors will be dis-
played virtually as accurate as possible
from Photoshop and the color engine
of Photoshop. It is helpful in most cases
to receive and evaluate a preview of the
expected results in print on the monitor.
One will refer to this kind of preview on
the monitor as a "Softproof".
Within the menu"Preview" choose the menu item "Softproof --> Cus-
tom...". In the following dialogue choose the profle, which describes
the subsequent printing process.
Tip: Most of the times it does make sense to choose the CMYK working
space selected from the Colorsettings at this point. This one resides at
the very top in the shortlist. "Relativ colorimetric" will be chosen as
standard rendering intent. For images with highly saturated colors it is
recommended to choose "Perceptive" in order to obtain the details in
these colors. The boxes for Blackpointcompensation and paper simula-
tion should be activated.
Attention: In case the images shal be printed in Offset- or rotogra-
vure it is urgently advisable to apply the standard profles from the
ECI-Homepage. The preinstalled profles in photoshop with names like
"Euroscale" and so on should not be used.
After approving the settings the preview can be toggled on and off
with a key shortcut (Mac: command-Y; Windows: Strg-Y)
In doing so the titlebar of the image window will show which color
mode the image is presented in(Attention: Only the color mode will
be displayed), how the data is coded and which color space is simu-
lated.
Next to the visualisation it is also helpful to be able to recognize the
color space of the actual image at all times. For this Photoshop of-
fers the possibility to display the chosen profle for each particular
image in the information section / status bar. Therefore it is necessary
to switch to "Document profle" in the information section (by click-
ing the tiny arrow on the right hand side of the section). The newly
designe histogram palette is an additional essential tool of Photoshop
Image
Engineering
CS. This will not onlysow the histogram of the
fle, but additionally the colored histograms
of each individual color channel. The palette
can be accessed via the menu item "Window
--> Histogram". In the histogram setting (tiny
arrow in the upper right hand corner) the
display of a statistic can be shown as well as
activating the display of the single channels.
Image
Engineering
Separation of images
For certain jobs it can be necessary to completly the data for print. In
this case it is the photographers assignement to convert the RGB data
in ready for press color separations ("CMYK data"). (Attention: This
kind of conversion should only be performed on special requestfrom
the clients side. In case the images are supposed to get edited at some
later time, an RGB (ECI-RGB, ...)color space makes a lot more sense as
a transfer color space !!)
Formerly these separations were created with so called "Separationt-
ables". This technique has been replaced by the use of ICC-profles. It
must be pointed out that the quality of the separation (and thereby
the subsequent usability on the press) is primarily dependent on the
apllied profles. For this reason
it is explicitly refered to the
standard profles from the ECI
homepage at this point.
These profles will achieve a
top-quality result under all
sorts of printing conditions and
paper types. It is disadvised to
apply the profles Euroscale"
"Swop" "Japan" installed by
Photoshop. The separation
of the data can be initiated
via the menu item "Image -->
Mode --> convert to profle".
In the dialogue the designat-
ed color space will be chosen
and the rendering intent ap-
pointed (when applying the
"relative colorimetric" please
note that a great deal of de-
tails will be lost in highly satu-
rated colors). By activating the
"Preview" box the result can
be visulised before actually
processing the data. Once the
separation occured the image
data is available in CMYK color
mode.
Image
Engineering
In principal the following applies to all digital cameras: as in analogue
photography the illumination is dependent on these three factors like
photosensitivity in this case in regards to the sensor the exposure
time and the preset camera aperture.
ISO-sensitivities
With analogue cameras the photographer determines the sensitivity
through the choice of f lm material. With digital this isdifferent by
nature, since the recording medium is a CCD- orCMOS-sensor , that is
built into the camera.
Therefore the sensitivity is always dependent on the image sensor and
is indicated by ISO-steps just like with analogue flm material. In oppo-
site to f lm it can be changed which means that the photographer
can choose certain ISO settings and switch the camera from ISO 100 to
ISO 800 or signif cantly higher settings.
These changed settings should be subject to caution though, since the
physical basic sensitivity of the CCD- or CMOS-sensor will stay the same:
The higher values can be achieved via an electronic amplif cation of
the image signal . Subsequently this also implies the basic background
noise to get amplif ed as well. Many simple consumer cameras with
high ISO settings will not produce images acceptable for the professio-
nal beyond ISO 400. Premium digital SLR camera systems and of course
the professional digital backs will still produce very good image mate-
rial in many cases even beyond ISO 1.600 or even ISO 3.200. With these
images full of atmosphere can be produced even without the use of
specifc artifcial light though ideally suited for event photography
indoors or even shooting at night.
Studio- and f ashlights
When it comes to lighting for digital cameras there is
not even signif cant rethinking required. Professional
SLRs or and even digital backs will cooperate brilliant-
ly with common Studio f ashlights or a system fash-
light. Only high resolution scanning backs will partly
require a ficker-free steady light system (like HMI),
could mean trouble for a photographer.
Please keep in mind that current available digital ca-
meras offer a signifcantlysmaller dynamic range than
the rather well-tempered negativeflm material from
analogue photography. Therefore extreme differences
2 Photography
2.1 Exposure and lighting
Image
Engineering
in contrast can create problems, thus either "loose" the dark details in
the shadows or clip the highlights and loose the details in this parts of
the image. (see also "tip" on the next page).
A digital SLR is usually equipped with a x-synchro contact for trigge-
ring a studio fash system. All studio fash systems can e triggered that
are equipped with this kind of contact. Naturally remote transmitters
via infrared or other ways to trigger a fash are possible.
Generally one needs to pay attention to the white balance setting
of the digital camera when shooting under artifcial light. Either the
camera will be manually adjusted in terms of color with a white piece
of paper or a greycard or a whitebalance preset for fashlight respec-
tively daylight needs to be chosen. Naturally the use of a designated
measuring device for the creation of individual color profles for the
specifc camera is recommended, in order to produce constant color
values. For the same reasons it is essential for a studio fash system to
deliver a power / fash intensity as constant as possible as well as con-
stant color temperature. Some studios will perform complete measu-
rings of the according fash systems, in order to utilize only those sys-
tems for the subsequent set assembly, that all feature matching color
temperature.
For a reportage assignment one is responsible for seeing that
the digital SLR and the according system fash can communi-
cate with each other: In many cases a regular TTL-measure-
ment for the fash intensity is NOT possible but one has to
fall back to the specifed TTL-technology (D-TTL, E-TTL) of
the camera respectively the matched upon system fash.
That's why many manufacturers offer new product lines of
system fashlights for their digital SLRs, that are adapted to
these special characteristics.
Aperture settings of the camera
The Aperture setting of a digital camera is similar to a regu-
lar analogue camera. In relation to the used lens maximum
and minimum aperture settings are allowed, referring to the size of
the opening in the lens that determines the amount of light falling
onto the CCD-sensor and in cooperation with the exposure time af-
fects the correct exposure of the image. With simple consumer came-
ras the aperture is often set via cursor keys in the settings menu on the
LCD-monitor, with a digital SLR the aperture setting can be achieved
defnitely more comfortably by means of a rotating knob or directly at
the aperture ring on the lens.
As a result of emerging defection effects with extremely small aper-
ture settings digital photographers are recommending f16 as mini-
mum aperture setting, even better with only f11.
Image
Engineering
Exposure - problems
Even with the seducing high image resolutions, which are so-
metimes even higher than those of SLR cameras many of them
are not appropriate for professional use because the exposure
clearance is too small for various reasons.
It starts with the lowest possible ISO-sensitivity for the camera,
which is simply too high for most tasks. Almost all models are
starting wit an ISO-value of 100 and some even at ISO 160: In
case of a studio-shooting this can mean that the fash system
can not regulated low enough and therefore one has to work
with simply too much light. This can be extremely fatal for ex-
ample with portrait photography for which a larger aperture
opening in order to outline from an out-of-focus background.
But in this case only a too small aperture settings and therefore too
much depth-of-feld will be applied with a consumer camera. The mi-
nimum aperture setting with many consumer camera is only f8, be-
cause of the small measures of the mini-lens as well as the recording
sensor and subsequently the according overall geometry inside the
camera aperture settings beyond f 8 would result in defection ef-
fects (see also "Aperture settings")
Outdoor shootings with sunshine are therefore only possible with ex-
tremely short exposure settings and wont allow any motion effects by
dragging the camera in motion while taking the picture.
A neutral density flter is pretty much the only thing that will help to
tease out a couple of f-stops.
Tip: Underexposure is better than overexposure
High dynamic ranges images thus extreme contrast between light
and dark it is recommended set the exposure on the highlights and
to put up with an underexposure in the dark areas in an image. Since
the dark areas in an image can be retrieved by a contrast- or tonal
range optimization even though this is not visible on the monitor
at frst the information in the highlights can not be saved from the
completely white elements of an image in case of an overexposure in
the highlights.
In order to optimize the exposure the professionals will use the raw
data format most of the times, in order to develop the images "digi-
tally" systematically. Additionally the conversion of raw images to 16-
bit data is a big advantage in terms of postproduction, because a lot
more tonal gradation is available for editing the image in Photoshop
after the conversion into such a 16-bit tiff fle. There the dark portions
can be brightened selectively in order to receive a more balanced fnal
result.
Image
Engineering
Digital cameras save its images in the shape of digitally coded thus
converted to numerically values- fles. The formal design of these fles
can be very different. A couple of standard formats have accomplished
themselves which all have their advantages and disadvantages and
therefor coexist in digital photography.
Color depth
A fle of an image usually exists from the three basic fundamental
colors red green and blue. We describe the color depth as the bit-
quantity, with which these shades for each one of these colors can be
displayed. In 8-bit we can digitally save 2 to the power of 8, thus 256
different numerical values. For each individual pixel a brightness value
between 0 and 255 will be saved for each of the three RGB color chan-
nels. Every pixel can consist of 256 green, 256 red and 256 blue tonal
values thus from totalling 256 x 256 x 256 = 16,8 million colors.
Increasing the color depth will result in the ability to reproduce even
more shades. Many digital cameras for this reason work internally
with a signifcantly higher color depth, than will be saved at the end
of the capturing- and editing process in the 8-bit image fle. 12-bit are
quite common for example which process 4.096 tonal value shades per
color excerpt, respectively even 14 bit, with which even 16.384 shades
can be achieved per color excerpt.
Even though the human eye is by far not capable to differentiate the
16,8 million colors of an 8-bit display, the higher color depth is im-
portant. The reason for this is the different brightness reception of
humans and digital cameras. The camera is able to differentiate very
well in the bright areas and the eye in the dark ones. With the larger
amount of shades and the human reception a better photo dynamic
can be achieved with the optimized gradation of brightness. The high-
er color depth is also used in image editing for additional contrast-
and tonal value optimization.
Compression
One of the biggest problems in digital photography consists i hav-
ing to save the immense amount of data on a storage medium. A 6-
mega-pixel image consists of altogether 6.291.456 pixel. For each pixel
the camera has to save the color information for red, green and blue.
Each color information consists of 3 x 8 bit (256 shades) per color (see
color depth), thus the image consists of 6.291.456 x 24-bit, which are
150.994.944 bit respectively 18.874.368 Byte (approx. 19 megabyte).
Storage media for digital cameras are not cheap though. For this rea-
son one was looking for possibilities to minimize the data. The solu-
2.2 Formats, Color depth and compression
Image
Engineering
tion was found in the "compression" of the fles, thus in mathemati-
cal procedures to save the data more space-saving. As a fundamental
principle we need to differentiate between two techniques:
losless compression (e.g. Tiff-LZW)
lossy compression (e.g. JPEG)
A lossless compression is characterized as a procedure in which the
original fle in spite of compression of the image can be reproduced
1:1. A simple lossless compression consists of saving redundancies (re-
peating information) in a concentrated form rather than singular.
The Tiff-LZW format applies this procedure but achieves only very low
compression ratios, since the pixels of a photo practically all hold very
different color values.
For this reason the Joint Photographic Experts Group proposed a new
compression technology and while at it even named it after itself:
JPEG. It takes advantage of the weakness of the human eye to differ-
entiate color differences less exact than brightness differences. JPEG
reduces by a mathematical procedure "discreet cosinus transforma-
tion DCT" the color information of the image by summarizing the col-
or information of multiple pixels to one information, the brightness
values will stay untouched. Due to the loss of some of the original
information the images will not be reproduced into the exact original
condition thus will be saved lossy. This loss is accepted because
a) the loss of quality with a cautious JPEG compression are not visible
and will stay tolerable even with a stronger compression.
b) the space saving of JPEG is immense: The initially mentioned 19-
megabyte fle of a typical photo will be reduced to just under 3-mega-
byte even with a weak JPEG-compression.
JPEG-format
JPEG has established itself as a standard with digital cameras. While
featuring a strong compression power it offers further benefts:
the actual intensity of compression is scale able. The more intense
the smaller the fles will get yet the quality loss will increase. Digi-
tal cameras offer the chance to between different JPEG grades. The
photographer gets to decide whether he prefers a smaller fle size or
better quality.
JPEG can be opened, edited and saved by practically every single im-
age editing application available. But it is not only for image editing
applications but also for the internet one of the most essential for-
mats for photos.
Graphic- and layout applications increasingly utilize JPEG-images
Many cameras will save the images directly as JPEG , which can be
essential due to the amount of storage necessary as well as the pro-
cessing speed. In postprocessing the photos shall be saved into the
Image
Engineering
JPEG format at the end of the processing due to the lossy compression
while saving.
TIFF
Tiff-fles are normally saved without compression. Every pixel contains
the complete information, provided by the camera.
This usually leads to large image fles. Tiff also supports 16-bits color
depth. This is an advantage since a lot more shades are available for
each color nuances. Practically every layout- or graphics application is
compatible with Tiff.
RAW-format
The raw formats take an exceptional position among the image fle
formats. The information coming from the recording sensor will be
saved without further compression or interpretation. As an outcome
the camera will not actually save a digital image in the truest sense of
it, but saves "raw data" on the memory card. A normal image edit-
ing application can not read this information because of saving the
raw fles and the according raw formats depends on the according
camera manufacturer. Thus the raw data can be manually edited to
every detail with the appropriate application, even the optimization
of whitepoint as well as the subsequent sharpening of the photo is
possible appropriate image editing application / an RAW-converter.
Since the possibilities of editing are so immense, the raw formats be-
long to the most popular fle formats in the felds of professional pho-
tography. Instead of leaving software sharpening, tonal correction via
Gamma-coontrol, behavior of colors as well as other parameters to
the interpretation of the camera, the photographer can convert ac-
cording to his own gusto with the accompanied or third-party conver-
sion tools. Adobe Photoshop offers since version 8.0/CS the possibility
to read many of the manufacturers own raw-variations. To open such
a raw-fle kind of resembles the development of an analogue flm in
a classic photo lab, with the opportunity to take infuence on the out-
come of the photos to a large extent.. For this reason more and more
consumer digital cameras feature the digital raw format.
By this the color depth of the recording system can be obtained and
will not get reduced to 8-bit. This guarantees a for higher dynamic
range in the photos.
The main disadvantages of raw fles lies within the slightly more com-
plicated and thus time-consuming handling because the photos have
to be processed compellingly, in order to get displayed and used fur-
ther on the computer.
Image
Engineering
DNG
Adobe is fond to replace the specifc manufacturers raw data formats
with a consistent "digital negative" and put a proposal on the table
during Photokina. This kind of development is eligible from the users
perspective, whether the manufacturers will jump on it needs to be
awaited. The DNG is an enhancement of the in ISO 12234-2 standard-
ized Tiff-EP (electronic picture).
JPEG2000
An interesting, but frankly in digital photography so far not a very
widespread format is JPEG2000. It does offer a lot of advantages like
e.g. the losless compression and high quality with strong compression.
Owners of Photoshop CS get the chance to read and write JPEG2000
fles The optional plugin is actually application CD under goodies.
Tip:
It is advisable whenever it needs to go fast and an individual postpro-
duction is not possible, to choose JPEG with a low compression = high
quality. But if it is possible in diffcult situations to save in the raw data
format, then it should be applied. Additionally a subsequent control
of the histogram is recommended directly after the exposure (if the
camera makes this possible).
Some of the modern digital SLRs are capable of saving raw data and
JPEG-fles simultaneously on the memory card.
Links
www.jpeg.org/
Image
Engineering
The term resolution is often equated with the number of pixels (actu-
ally referred to a sample rate), that a camera delivers. This equation is
only applicable under ideal reproduction conditions, from which real
cameras are more or less far-off. A bad lens in front of the camera
means that even with an extremely high number of pixels not a good
an detail-rich photo can be captured. The quality of the reproduction
is dependent on the camera and lens used and hence con not bet con-
ducted any further at this point. The following observations will thus
relate to the idealized reproduction in the shape of pixels.
How big of an image fle is necessary in order to get printed precisely?
Every photographer should ask himself this question, before sending
his photos to the printer. Because only if there is suff cient enough
data available for all the details of the motive, they can get repro-
duced perfectly. Especially with illustration books, magazine photos
or further large formats one can accept no compromises at this point.
A careful planning process already starts before the actual capture,
since there is no way to add the data afterwards. For instance chang-
ing the size afterwards- the so called interpolation will not increase
the resolution in terms of detail reproduction. Since this will create ad-
ditional pixels purely through average determination the photo will
come across dull and out of focus after the operation. For this reason
the right setting needs to be chosen, even before the shutter release
button is pressed.
Resolution on screen and on paper
Just how the image format affects the
f le size, is describe in the previous
chapter. However kilo- and megabytes
say little about to which degree a pho-
to can be printed on paper in a high
quality.
This is particularly affected by the print
format and method. The number of
pixels depends on the specifcations,
from which the exposure needs to con-
sist of marginally.
These dots are additionally referred to as pixels, an artifcial english
word from picture and element. The pixel as a fundamental element
is no constant dimension, but is very different depending on each out-
put medium. Thus they can be quadratic or squarish depending on the
2.3 Resolution
Image
Engineering
resolution or rather tiny colored smudges, that can be on top of each
other in contrary to the monitor.
There is a basic rule to this: The more pixels are available the fner
the image can be dissolved. Yet the transfer of image data sets a limit
from which oversized amount of data makes no sense and cannot be
handled anymore.
A good resolution is achieved if the pixels create the impression of
constant color and brightness gradients, without the single elements
becoming visual. On the monitor 96 dpi can comply with these require-
ments on the distance of an inch (ppi), without the ye recognizing the
composition (one inch is equivalent to 2,54 cm). On a piece of paper
there have to be even more colored pixels on this distance, because a
photo or an image will be observed mostly from a shorter distance. A
printer for example should be able to produce approx. 300 "real co-
lored" print dots per inch (dpi) in order for colors and shapes to come
out clearly. Wanting to reproduce a photo on a piece of paper in the
same size as on the monitor it must contain theoretically approx. nine
(3x3, width x height) times as many pixels.
The printing screen will dissect such a "real colored" pixel.
Only Thermosubliminations printer, autochrome printers or image set-
ters have the capbility to create a printed dot for each pixel in the
appropriate color. These can suffce with just 150 dpi if necessary in
order to receive very good results.
An inkjet printer or an offset machine
is technically not capable of producing
such continuous color gradients. The so
called Iris-prints are exempted, that ac-
tually operate with "continuous tone"
thus continuous color gradients. A regu-
lar inkjet printer though has to place the
basic colors in such a way on top of each
other, in order to simulate a merged color
to the eye. A purple smudge on closer ex-
amination thus consists of a collection of
dots in magenta and cyan, crowding itself
in a so called "screen cell".
These screen cells match the according
print dots. A further unit referring to re-
solution is: lines per inch (lpi). Nobody is
actually still working with "lines" this nomenclature is actually from
early printing days.
Thus one can fnd in a newspaper photo with 100 lpi 100 screen cells
on one inch. These screen cells themselves consist from a multitude of
Image
Engineering
smaller dots. The offset print with a screen is the regular process for
newspapers, magazines, books and editorials. Subsequently you will
fnd the according resolutions for each print item. If necessary one
needs to ask the editorial offce or the printer itself.
Newspapers: 65 to 100 lpi
Books (uncoated paper): 120 to 133 lpi
Books (coated paper): 133 to 150 lpi
Magazines: 150 to 175 lpi
Art books and -magazines: 175 to 250 lpi
Thermosubliminations printer: 206 to 400 dpi
Practical example calculated
Back to the photographer and his photo: The measurements with digi-
tal cameras are indicated in absolute pixel values approx. 1600x1200
pixel. Additionally they could be specifed in ppi (Pixel per inch).
Misleadingly the specifcation in dpi (Dots per inch) has achieved ac-
ceptance as a term for image resolution. And thus creates confusion,
since dpi only refers to the actual printed dots, with which the pixel of
the digital image will be processed.
According to the utilized printing screen the following formula can be
applied, as long as its specifed in lines per cm:
screen width (cm) x 2,54 = screen width (lpi)
To play it safe a quality factor is added for the calculation of dots
per inch, which is ideally specifed as 2. In practice due to the human
reception values between 1.4 and 1.6 have been established. So the
formula sounds like this:
screen width (lpi) x quality factor 1,6 = dpi
as an example we will perform this calculation for the monitor dis-
play:
A photo has the dimensions 800 pixel times 600 pixels, thus altogether
192.000 pixels. Every pixel of the fle on the screen will be displayed as
a pixel on the monitor, which usually operates from 72 dpi to 96 ppi.
The result is a width of the photos is 8,3 inch (800 divided by 96 ppi)
or 21,2 centimeter.
Because many beginners believe, that each pixel of the camera is in ac-
cordance with each printed dot from the inkjet printer the following
error originated: One thinks the printer will print the image with
an resolution of 1.440 dpi. This would result in an image with 800 x
Image
Engineering
600 pixel that will be printed with a width of approx. 1,4 centimeters
(800/1.440 x 2,54) and a height of 1,05 centimeters.
Frankly this assumption is wrong due to the above described rasterisa
tion process: Since the output system can only assemble the digital
pixel only through a screen / sample from different printed dots, in
order to create millions of shades from its original colors, it has to fea-
ture a resolution multiple times higher than the image resolution of
200 to 300 ppi. These 200 to 300 ppi will then get printed with 1.440
dpi. The screen of an inkjet printer usually will not create so called
screen cells, as in the common offset printing process, but utilizes sta-
tistically spreaded dots. It is "frequency modulated".
Color- and greyscale printing screen
A screen cell can be flled with colored or black dots. According to its
density and composition the outcome are either shades of grey or a
certain tone of color.
With a black & white print for example the screen cell could consist
of a matrix with 8 times 8 printed black dots. Form this we can ac-
complish 64 shades from white (empty cell) to black (flled cell). In
a professional production print most of the times 4 colors are used:
Cyan, magenta, yellow and black, CMYK if abbreviated. These are also
commonly utilized in regular inkjet printers.
As shown in this screenshot the different resolution settings can be de-
fned in the settings menu of almost any digital camera. If one knows
what size the photo will be printed, one could make an exact choice:
(printwidth in centimeters / 2,54) x resulting dpi = necessary image
width in pixel.
The height is a result of the exact as-
pect ratio and does not need to be
calculated separately. Frankly one will
seldomly meet the exact image
dimension like this and thus will rather
acquire the dimensions, in which the
available image material ca be printed.
(width in pixel / acquired dpi) x 2,54 =
printable width in centimeter.
(height in pixel / acquired dpi) x 2,54 =
printable height in centimeters
Tips when photographing and exposing
At last a couple of considerations when shooting: The pixels are not
always based on the actual information of the hardware. Some came-
ras promise high resolutions by adding calculated pixels. This process is
called "interpolation". Since this method is based on approximation,
Image
Engineering
a certain out-of-focus and dull colors will often creep in. Thus the in-
terpolated resolution can be ignored and only the optically available
pixels should be used.
The same is true for the use of the so called digital zoom. Instead of
enlarging an image scene with a powerful lens, the digital zoom func-
tionality will blow it up with calculated pixels. The consequences are
out-of-focus images and reduced contrast in this case also. As a matter
of fact there is no reason to worry about insuffcient optical achie-
vable pixels these days. In the meantime almost every digital cameras
comes with at least 3 megapixel, 4 to 5 megapixel are standard, pre-
miumconsumer cameras capture the images with 6 to 8 megapixel and
certainly solutions exist that can go beyond 11 and 22 megapixel.
Similar rules apply to having the digital images printed at your local
camera store as to the above mentioned thermosubliminations prin-
ter, because an output process is adopted, that can control color for
each individual pixel and it is not necessary to raster. An image resolu-
tion of 200 dpi is suffcient enough as image resolution.
Most photo dealers or online print services will state the necessary
resolution for the desired image
size mostly in absolute pixel dimen-
sions, because that matches the in-
formation coming from the digital
camera. For a 4x5 inch (10x15cm)
print a resolution of 1.600 x 1.200
pixel is recommended.
Links
www.agfanet.com/de/cafe/photocourse/beginner/cont_index.
php3?lesson_id=1&page_id=5
www.agfanet.com/de/cafe/photocourse/digicourse/lesson4/cont_
chapter05.php3
www.dpreview.com/learn/?/Glossary/Digital_Imaging/Resolution_
01.htm
Image
Engineering
When shooting a digital photo not only lens sys-
tem and chips take part, but numerous calcula-
tions will be performed, until the image is actu-
ally created: The light captured by the sensor as
voltage will be transformed into a digital signal
, the color temperature will be applied accord-
ing to the menu settings and the image will get
converted into an image format like JPEG (see
also 4.1).
The image will get sharpened on this way also.
The degree of internal postproduction can be
determined via a menu, but in most cases only
as a rather rough choice between the three set-
tings "soft", "normal" or "hard". The electronic sharpening has noth-
ing in common with the focussing of an image. A wiggly image due to
a long exposure time will not turn sharp because of it. In fact it is sup-
posed to compensate this kind of
out-of-focus due to technical rea-
sons, caused by the interpolations
of colors or resolution from the
camera. A similar technology is also
applied when scanning images.
The description in the settings menu
says very little about the actual re-
sults, since the chosen sharpness
very much depends on the calcula-
tion methods by the manufacturers.
As shown in the screenshot, edges
can be emphasized by very differ-
ent methods. The frst untreated
example shows a relatively wide transition from black to white, which
creates an out-of-focus impression. In the second screenshot the area
is crushed, by strongly increasing the bright-dark-contrasts one
method of sharpening. There is a risk though for color gradients to
get torn up. With the "unsharp masking" method (USM) only small
sections will be considered.
In the extreme case as you can see here sharpening will create sig-
nifcant fringing. Particularly with enlarged displays will annoy just as
much as out-of-focus.
Every manufacturer applies his own defnition of "normal" or "hard".
Even the setting "soft" will not be left out from an internal sharpen-
ing during interpolation.
The illustration is showing three diferent versi-
ons of Sharpness settins, created from the same
cameras. The "soft" setting is shown on the left,
"normal" in the middle and the results of the
"hard" setting is shown on the right.
Many digital cameras offer to
choose between different Sharpness
settings via the in-camera menu.
2.4 In-camera sharpening
Image
Engineering
There is the advantage of in-camera sharpening since the sharpening
algorithm is adapted to the "contrast transfer functionality" of the
camera. This actually functions better than e.g. Photoshop algorithms.
A soft cautious sharpening from the camera can be recommended in
many cases.
Many photographers choose a softer setting on purpose since the elec-
tronic sharpening can be performed just as well in an image editing
application. The degree of sharpness via USM-flter can get adapted
exactly to the individual image because the flter impact needs to be
adopted exactly to the resolution of the image, the printing process
and the dimensions of the subsequent print.
Besides simple flters that are part of the standard equipment of im-
age editing applications in which the user can manually adjust the
USM-parameters there are also some third-party flters specialized on
digital sharpening processes, that can perform the necessary settings
via automatic or semi-automatic. The sharpening automatic could as
a special service also incorporate vignetting with wide angle lenses as
well as increases noise with high ISO-settings. But mostly small digital
cameras for example will apply stronger sharpening, in order to com-
pensate for shortcomings in the lens and low resolution.
Otherwise it is up to the photographer whether he trusts the in-cam-
era software rather than the visual handling. At least increased noise
in the images could be prevented b manual sharpening with a couple
of tricks, but the internal processing will write it into the photos ir-
reversibly.
Image
Engineering
The term "color space" is occupied by multiple meanings and thus it
leads to some confusion.
Colors can be describe in different mathematical and physical three-
dimensional models. Thus the term "color space". The "color models"
Lab, RGB, CMYK are part of these scalings.
The RGB or CMYK values from a certain device like a camera or moni-
tor can be visualized within the Lab presentation, which is why it is
also occupied with the term "device color space". This is also true for
idealized working spaces like eciRGB or ISO-Coated (CMYK).
Digital cameras, monitors and Scanner use the additive color model
for displaying the individual color shades, in which the three primary
colors red, green and blue create the different combination colors.
With maximum brightness these three colors RGB produce white in
this combination, black in the opposite, if none of the three colors is
visible.
The RGB model is equivalent to our vision, because in our eye these
three primary colors are received by the photo receptors responsible
for the color impression and our brain combines this color im-
pression by mixing these signals.
The subtractive color model assumes white light, from which the
colors cyan, magenta and yellow will be withdrawn as a spec-
tral part. This way cyan withdraws red from the light, magenta
the green and yellow the blue. All printed matters whether
they were created with an inkjet printer or in a printing plant
- have to utilize this model. With them the combination of all
colors will not result in white, but black if color is absent the
paper can be describes as white.
RGB and CMY are tightly linked to each other. If two primary colors
from one model are combined a primary color from the other sys-
tem is generated.: red and blue will emerge magenta, red and green
will emerge yellow and the combination of green and blue emerges
cyan.
In theory the conversion from color model to the other would seem
rather simple. In a complete color model all colors from the additive
system RGB can be converted to CMY and vice versa. The two graphs
make this quite clear.: If they are displayed in the HSB color model
which describes colors as shades of color, brightness and saturation
values, the primary colors of RGB and CMY are exactly opposite of
each other. Red is the opposite of cyan, yellow the opposite of blue
and green the opposite of magenta. If you turn the color wheel by
180 degrees than the colors will get relocated to the prior position of
their opposite. Cyan is now located at the position of red, yellow on
2.5 Color space
Image
Engineering
the position of blue and so on.
Frankly in practice it is not that simple. One of the problems
with the CMY-color model is based on the situation under full
color coverage of the three primary colors it is supposed to be-
come black, but this black will not be achieved this way. Since
the printing inks can not be manufactured pure enough , it cre-
ates a mixture of a dirty brown-green. In order to actually print
black it has to be added as a fourth color (K for blac-K, Key
or K-ontrast). For the calculation of this black contingent are
two different methods available: With the UCR - under color removal
that hold the same three printing colors to the same degree and thus
will create greyshades as a result of the composition, will be replaced
by the fourth color black. The degree of these greyshades, which are
describe by a gradient curve, can be adjusted in an application spe-
cialised on this kind of separation.
The grey component replacement (GCR) on the other hand reduces all
colors by a certain degree and rebuilds it by adding the fourth color
black. Just like it the grey component replacement is described by the
gradient curve in a separation application. Since there is no way to tell
from an CMYK image, how it was built, a recalculation to RGB-mode
is not possible in an exact fashion. Additionally when converting from
RGB to CMYK another factor is applied during calculation, which de-
scribes the so-called dot gain. A very absorbent paper - as it is the case
with for example newspapers the inks will scutter and create larger
print dots, which will infuence the theoretically correct color- respec-
tively brightness value. This factor can not get recognized either in
the CMYK-fle at a later point and thus aggravates the reconversion
to RGB.
In practice these two individual color models feature a different range
of color. In other words, the digital camera or the RGB display com-
puter monitor can display numerous colors that can not be produced
on a CMYK printer. Those colors beyond the reproduction capabilities
of a CMYK printer need to get converted to those reproducible by
CMYK. A pretty good example are highly saturated green tones or
even strong orange tones. The according conversion is very complex
and offers a couple of methods of resolution. For instance the maxi-
mum feasible color of the CMYK destination color space can be used
as a replacement for the non-printable nuance (colorimetric render-
ing). Alternatively the complete RGB-color space will be compressed
highly, in order to ft all of the non-printable back into the destination
color space (perceptual rendering intent).
Color management system will help with the conversion. These consist
Image
Engineering
among others of so called "ICC-profles", their design was defned by
the International Color Consortium. ICC-profles describe the differing
color behavior from a reference curve of any device. Profles for input-
as well as output devices can be defned, whereas photographers do
have a special interest on profles for digital cameras and scanners.
Most digital cameras will work in the sRGB color space, which already
contains a large portion of the visible colors and is quite frankly still
larger than the RGB-color space which most monitor will display.
Just how the exact sRGB-defnition for the according cameras will look,
is still the manufacturers secret, because even with "standard" sRGB
there are signifcant differences arising. Thus it is recommended to cre-
ate an individual profle for the according camera (see also 1.2). The
color management system will then "know" which color space can be
captured by the camera. The photos created with this profle will now
have to get converted to an output color space for output purposes.
Associations like the german BVDM, FOGRA and UGRA make color
profles available for certain printing processes, that are constantly
under development among these the ICC profles ISOcoated.ICC and
ISOuncoated.ICC for offset prints on coated and uncoated paper.
It is important for the photographer, that he will merely be able to
display the CMYK on his monitor in a limited manner.. The DVSplus
system was developed for this reason , which consists of ICC-profles
and reference prints. The whole purpose of this technology is to offer
the photographer a display on his monitor for his RGB-data , which
is limited to the printable colors in an offset process. They are RGB-
profles that will practically allow a secured CMYK-display as a proof
on the monitor. They are especially recognized with highly saturated
colors in the image. The could lead to arguments between the pho-
tographer and the client, because they can be seen on the monitor in
brilliant colors, but they will not be recognized on the
sequent print result.
In other words: With DVSplus the image can be reduced to those col-
ors that any printer is able to put on the paper. Thus the profles can
be used for CMYK separation. A controlled color space compression
will be performed when applying the DVSplus profle, that will ft the
Rgb color space of the digitally captured image to the signifcantly
smaller Offset color space.
Links
www.fogra.org/
www.ugra.ch/
www.bvdm-online.de/
Image
Engineering
ImageEngineering
Augustinuustr. 9d
50226 Frechen
Germany
Phone. (0 22 34) 91 21 41
Fax (0 22 34) 91 21 42
www.image-engineering.de
info@image-engineering.de
Digital camera tests
We test digital cameras and scanners for magazines like the German
Color Foto and ct since 1997. In the mean time numerous manufac-
turers also use our services to test prototypes of their cameras. We
also offer consultance as well as complete test stands including test
charts, illuminators, and evaluation software.
Trainings
You can proft from our experiance in camera testing and practical
image processing by the trainings we provide. We offer a wide range
of courses from the basics of image processing up to the use of a
digital camera as a luminance meter. The image enhancement for
internet purposes and colormanagement aspects are also part of the
program.
Digitalization
We digitize prints, slides, negatives, old books and historic doku-
ments. If you want to do this yourself we offer consultance and help
you to avoid the expensive mistakes we made in the past. We have a
lot of experiance in handling rare and fragile originals in combination
with optimized performance.
Mass digitization of 35mm slides
Image Engineering offers a cost-effective high speed solution for the
digitalisation of slides. A slide projector is modifed for this purpose so
that the slide is diffused and uniformly lit. A digital SLR camera which
is ftted with a good macro lens, looks from the front into the projector
and photographs the slide. Interested? Look at our website.
Imaging is our profession
Image
Engineering
Images photographed with digital cameras can have strong differences
in color according to the time of the day. At noon the sunlight consists
of lesser portions yellow and red, but lies within the bluegreenish
range. In the evening though the red spectrum will prevail. The same
is true when capturing under lucent orange light bulbs or blueish
fuorescent light tubes a clearly visible color cast will be seen.
In order to prevent the unnatural coloring, the camera needs to be
adjusted via white balance to the according light. In the analogue
photography such flms were used which chemical emulsion would
react stronger or weaker to parts of the spectrum. The method is
endorsed by the use of conversion flters. These colored lens cantilevers
would f lter the unwanted components, by letting especially the
complimentary colors pass. In digital photography, either the white
balance automatic can be used, the manual white balance settings
via menu settings or the images will be edited in the computer
afterwards.
The spectral composition of the light is def ned by the so called color
temperature, measured in Kelvin. Thus a theoretic model of an ideal
black body is presumed, a black hollow balls with a hole in it. It simply
swallows all incident light, but will emit red- to white glowing light
according a heating through the hole. The rising sun thus complies
with approximately 5.00 degrees kelvin, a clear blue sky complies with
11.00 degrees kelvin.
The eye will not recognize these kind of
fuctuations, since the brain will partly
compensate for it. Thus the camera needs to be
adjusted to it via white balance. A blank sheet
of paper should therefor appear white in any
kind of situation, no matter when and where it
was photographed.
Almost all digital cameras offer besides the
white balance automatic predetermined presets
to choose from, like "daylight" or "tungsten".
Admittedly they will only deliver approximated
values, the actual color temperature will not get captured. Even if
acceptable results can be achieved by this, method will f nd its limits
with mixed light situations. This will always result in a color cast. The
most exact white balance can only be achieved via a manual setting
to the according color temperature, as long as this functionality is
available is the camera. A sheet of white paper or a greycard can
serve as reference according to the camera. The only important
thing in terms of the camera is not to have a color cast on it, the
luminance value of it is not important. The complete color nuances
2.6 White balance
Image
Engineering
will arrange themselves to this neutral values. If the white balance
was not performed correct during capture, but the color discrepancies
are not too trong, than it can be corrected within an image editing
application without problems. For raw data, which are
saved unaltered by the camera software to a storage
medium, this is actually the standard procedure.
Image
Engineering
In-camera memory
There are a bunch of different types of memory cards, that have
evolved around digital photography. Only a few though actually have
a meaning in professional photography now. Generally speaking the
photographer should not buy one large single card but rather a few
cards with medium storage capacity in order to minimize the risk of
data-loss with a broken card. If a card should indicate an error while
shooting simply put it away safely in the frst place. In many cases the
data can be restored with a "Recovery" software. (e.g. JPEGdump).
CompactFlash
The memory card suited for professional purposes is the CompactFlash
card, not meaning the IBM Microdrive. They are mechanically to frag-
ile and should be used in professional photography. These cards score
capacities of up to 4GB meanwhile.
SD Cards
An alternative offered in some cameras is the SD card. It is rather stur-
dy and can obtain capacities of to 1 GB. It also features a mechanical
write protection.
All others memory cards like the MemoryStick, Multimedia-Card, XD
Picture Card or the SmartMedia are without meaning for professional
photography.
Memory for transfer and archival storage
The service providers in the production process are prepared for dif-
ferent transfer technology. Next to the electronic transfer via ISDN
(see also "5.2 transfer technology") the CD-R has been established as a
lowcost standard for the offine transfer of large amounts of data.
CD
So the transfer of large images and layout fles takes place via one-
time writable (CD-R) and rewritable (CD-RW) CompactDiscs. They are
rather sturdy since they get along well without any mechanical parts
inside. And if the chemical coating wont get damaged by either heat,
UV-light or solvent, the CD-R can last up to 30 years, the CD-RW will
last up to a year if used on a regular basis. It is commonly saved in data
format ISO9660, an accepted norm by every operating system. A CD
fts up to 800 mb.
2.7 Storage media
Image
Engineering
DVD
The Digital-Versatile-Disc thus can save up to 4,7 GB with one layer of
data (DVD category 5). It is commonly written in the Universal Disc
Format. Other formats like ISO 9669 are theoretically possible, but is
not readable by the reading devices. But the most popular drawback
are the competing DVD-Standards. The one-time writable (DVD-R) or
rewritable (DVD-RW) minus-standards follow the offcial DVD-consor-
tium, DVD+R and DVD+RW are vendor specifc.
So when buying pay attention for the blank to be compatible with
the DVD-burner and for the printers DVD-drive to be compatible with
the according standard. Those drives that are compatible with both
kinds should be preferred. The old standard DVD-Ram had a tough
start. Blanks of this kind could be rewritten more often than its rela-
tives, but there is no software necessary compared to common CDs
and DVDs to burn them. You can simply drag the fles to the drive in
the Windows Explorer. A DVD-Ram also offers the highest data secu-
rity available.
The development will not be fnished for long. Besides the already
available Double Layer-DVDs with 9,5 gigabytes the so called Blue Ray-
Discs are already announced, that will feature even higher capacities.
Those will be able to save up to 30 gigabytes on the same dimensions
as the CD or DVD.
Links
en.wikipedia.org/wiki/CD-R
www.informatik.hu-berlin.de/~hstamm/mo.html
Image
Engineering
Dust and fuzz are the enemy of digital photography. All stops were
pulled against the annoying teasers when enlarging flm: Compressed
air and antistatic appliances paired with a set of special gloves and a
special fastidious treatment of sensitive slides, negatives or prints.
When scanning not only the original but also the glass plate needs to
stay free of dust and fngerprints at all times, in order not to degrade
yourself to a retoucher with laborious desmudging. But even in the
camera the dust will raise the dust.
Cleaning the sensor
It used to be the flm on which dust and fuzz would settle down,
and was then simply replaced, the sensor is becoming the permanent
dust catcher - especially if all doors are open for dust after changing
lenses repeatedly. In addition to the tiny airborne particles, that my
be removed just as well through the classic cleaning methods just like
from the analogue camera ( a simple bellow is recommended yet the
cheapest method is a cluster ball but without talcum), smallest micro
particles will stick to the sensor due to air humidity and a breath of
lubricant.
Thus a simple blowout wont do the job anymore. The use of Q-tips
is actually forbidden since there is the danger of tiny fbres getting
stuck and affecting the mechanics. Chemically treated cleaning cloths
for glasses must not be used either neither for lenses nor for the
LCD, since the coating of the lenses respectively the display can get
scratched. There is also a strong disadvise not to use canned air ei-
ther, since some photographers underestimated the temperature-and
greenhouse gases problem. There are some rumors that lenses should
only be changed with the camera switched off, in order not to "ex-
pose" the sensor when "loaded" , which attracts the particles faster
and for fast shooting series to load up the sensor with static energy
theories with no evidence.
Eclipse is a liquid and Sensor Schwab are special cleaning cloths from
Photographic solutions in the USA, which are designed by size spe-
cifcally to the sensor (e.g. for Fuji S1 and S2, Kodak DCS, Nikon D1,
Canon EOS 20D) and recommended by Kodak, Leica and Fuji for the
cleaning of their cameras. The cleaning cloth should be customized to
the size of the sensors and used actually just once from each side. The
liquid is a high-purity methanol (use in constantly vented rooms only)
which also evaporates completely without residue, while the cleaning
cloth are produced and sealed in the clean room and will not fuzz or
scratch.
A test shot is mandatory after cleaning: attach a lens, switch autofocus
2.8 Camera maintenance
Image
Engineering
to off and set the aperture to 11 or 16 while shooting a white piece of
cardboard (preferably with an overexposure by 2 f-stops). The result
may still show some delicate remainders of dust particles when zoom-
ing in the image editing application, which will never become visible
though in a real photograph. The ambitious photographer should not
carry the cleaning into extremes due to micro sized dust particles: the
zooming steps of image editing applications have boosted the de-
mands of many photographers.
Even though the sensor is covered by a glass anti-aliasing flter or opti-
cal glass and the cleaning process is just about the same as cleaning a
lens, the cleaning of the sensor can become a fne motor challenge. If
you do prefer not to be exposed to the danger to damage the surface
of the flter, the lantern slide or the mirror by making a wrong move
or using the wrong cleaning device, can still call upon the regular
manufacturers service department.
It is almost always the blue sky that
brings out the little dirty secret. Dust
and fuzz on the sensor affect the
image.
The high zoom level
will show the artefacts,
caused by dust and fuzz
on the sensor.
Image
Engineering
Shake it, sensor ...
Olympus offers active dust reduction with the E-1 by a so called Su-
personic Wave Filter, which is attached in front of the CCD and makes
sure that dust and dirt particles will be waved off and caught durably
by a specially coated foil. Each time the camera is switched on this
functionality starts to work without any noise or vibration and re-
moves any dust and dirt particles at least as effcient as any bellow,
but more gentle and reliable. The service is rec-
ommended for fxed particles and exchanging the
dust catching foil regularly.
In the glittering light of the hot desert sun
Permanently damaged pixel can occur if the camera is aimed for a
long time at an extreme light source with the Aperture all the way
open. This will not happen with an digital SLR but the digital compact
keeps the aperture opened in order to "feed" the display with a Live
preview. But there is no need to start getting "chicken-livered" during
the next sunset: As long as the camera on a tripod is not aimed at the
glowing midday sun for minutes the sunset
may be shot full of relish during all phases.
In fog, wind and weather
The digital camera can be protected from
Afraid of pixel errors ? They can occur with con-
sumer digital cameras and camcorders, if they are
directed at a glistening lightsource for minutes. A
sunset wont harm no pixels on a sensitive sensor.
The supersonic wave flter is assembled in front ofthe
CCD. The ultrasonic waves produced by it are so fast
and effective that accumulated dust and other dirt
particles will fall off and get stuck to a specially coated
foil. This function takes place each time the camera is
started but can be activated manually too.
Image: Olympus
Image
Engineering
rain, snow or splash water with a plastic bag (Tip: the cooling bag
from the frozen-food department in the supermarket) with an open-
ing for the lens, because the electronics inside the camera are more
than sensitive to humidity. A custom tailored protective case is safer
in any case. Once you are back in your warmed-up home the camera
needs to be dried off with a soft and dry piece of cloth. For digital SLR
and compact cameras with a flter thread a uv-flter is the perfect pro-
tection for the lens just like the analogue camera. It is no big news of
sand and saltwater to damage the body and for that matter is true for
either analogue as well as digital cameras. If the digital camera should
have been exposed to bad weather conditions or a long foggy morn-
ing, just switch off the camera and remove the batteries and memory
card in order to dry it on the air for at least 24 hours.
On the way during frosty conditions
Frankly it is not only big heat and high humidity that get to the digital
camera while the LCD starts to slow down at low temperatures, the
camera may even deny service shortly during frosty temperatures. If
it gets truly frosty for a permanent period of time only a purely me-
chanical cameras will perform it is service reliably. If you are travelling
by car with your camera, simply grant your camera bag a place inside
the car, instead of keeping it in the cold trunk. Low temperatures will
also get to the batteries and the consequences can range from a slight
interference of power to a complete loss of power. The spare battery
better spend some time inside a small container of either a jacket- or
the pocket of your pants.
For your fngers to stay fexible even during very low temperatures, a
gel pack inside your coat pocket is the perfect companion on a cold
day. It contains a small metal plate, which activates the gel pack if
folded and unleashes its warmth. But only once its not warm enough
for the fngers anymore, it can be stored inside the camera bag. Digital
cameras, batteries and rechargeable batteries shouldnt be exposed to
temperatures over 40 degrees celsius for a long period of time.
Air trip
The typical question if travelling by plane: Is there a chance for the
digital camera to get damaged when x-rayed during security-check on
the airport. Will the metal detector damage or delete my data on the
memory card ?
But that is of no danger. Even though it is theoretically possible to
erase a storage position via strong ionizing radiation, but the appro-
priate power to do so is a loth higher than the power of the x-ray
machines during security-check. However if you feel that the x-ray ma-
chines during security check-up are not up-to-date when traveling to
Image
Engineering
the more distanced areas, simply remove the memory card and have
it checked manually
If travelling with a digital SLR and would like to take the above de-
scribed cleaning-set, you will be shaken down. The cleaning fuid is
highly fammable and must not go on the airplane.
Apropos memory cards
The exposed contacts from the memory card like smart media are their
achilles heel: they are vulnerable to corrosion and pollution. Rubbing
with a regular cloth may result in static charge. By which voltage will
get inside the card and fbres may dissolve from the cloth.
CompactFlash- and MemorySticks dont have contacts being exposed
and because of that they offer more data security.
Theoretically the CompactFlash- and SD-modules as
well as the Memory Sticks are increasingly resistant
to bit-errors than SmartMedia cards due to the built-
in forward error correction. Microdrives are rather
popular due to the extremely high capacity and
comparably low price. They are true small harddrives
though with a mechanic and vulnerable to crushes.
In favor for security they should rather remain in-
side the camera and should be read via the USB- or
frewire interface.
Source of supply
The cleaning fuid Eclipse from Photographic Solu-
tions and the Sensor Schwabs, that are recommend-
ed by many camera manufacturers and in numerous
photography forums, can be purchased from any
photo accessories specialist.
Links:
Field report sensor cleaning:
http://nikondigital.org/articles/ccd_cleaning.htm
http://www.bythom.com/cleaning.htm
http://www.photo.net/equipment/digital/sensorcleaning/
The userguides of me-
mory cards with exposed
contacts sent out explicit
"thumbprint" warnings.
They should be handled
by the plastic parts only.
Image
Engineering
There are those kind of pictures that are pretty self-explanatory. Thus
nobody will fnd it hard to attribute a picture of the burning twin tow-
ers from September 11th. Our experience tells us, when the shot was
taken, what it shows an where it was taken. These are the essential
informations about a picture. But who actually remembers the exact
date of the nemesis of the titanic (April 14th, 1912) or the fring of the
"Hindenburg" Zeppelin (May 6th, 1937). The non-informed viewer is
probably not even able to recognize what these pictures are actually
showing. That is the reason why it is so important to add a bunch of
information so called metadata which enable the interpretation
and classifcation. There are two standards: Exif and IPTC, that apply
to the description of images and are supported by most camera and
software manufacturers. The data from both of these standards will
be saved directly into the image data and therefore wont get lost ac-
cidentally. Since most image databases also support these standards,
images can be imported automatically into numerous databases and
thus will be found through a query. Additionally there is still one more
standard, which is based on the programming language XML with the
name of "Dublin Core", frankly it actually saves the image informa-
tion in additional XML fles, that can get lost during transfer of the
images. That is why this standard only plays a noteworthy role in the
area of historic archives.
(Exchangeable image fle format for digital still cameras, www.exif.
org)
This standard is in charge of the "technical metadata". It specifes how
the technical information will be saved in the image. A part of this in-
formation are date, time, exposure correction, white balance, .... a list
of all the available EXIF tags can be found at the time of publication
under http://www.awaresystems.be/imaging/tiff/tifftags/privateifd/
exif.html and goes all the way to GPS positioning data., if the camera
features a GPS receiver or the possibility to connect to one. Which
data will be entered to the Exif header, is dependent from the accord-
ing camera respectively the manufacturer. The manufacturer specifc
felds can only be read by the software to the camera.
Exif data is dedicated to record the technical capturing conditions.
Thus they can not be edited formally. A programmer will not have
a hard time changing it though, which is why they can not be used
offhand as a copyright- or modifcation protection. In case of a tech-
nical problem of the camera or the assignment of an image the Exif
3. Data optimization
3.1 EXIF
Image
Engineering
data often can add to a solution. That why it is important that the im-
age editing application will save the Exif data rather than removing it
when saving the image.
(International Press and Telecommunications Council)
IPTC information is all about a so called describing metadata. Which
means the IPTC data will describe where the image was taken and by
whom and respectively who owns the image rights, besides the image
content. These data wont be added automatically but rather need to
get entered either by the photographer or another person. Some im-
age viewers of camera manufacturers and also special "IPTC writers"
allow the semiautomatic batch entering of the IPTC-felds for an im-
age series. Photoshop supports the majority of the IPTC-feds via the
menu item fle information and allows the editing of IPTC felds, at
which not all felds are available all the time.
Please always put yourself into the position of the person viewing
your images, when entering the felds. Which questions will he ask,
The answers to these questions belong into the according felds.
The photo journalist needs to enter the following information in any
case:
Photographer: Who took the picture?
Credit: Who will get the fee?
Date: When was the photo taken?
Caption: Image description, who is pictured: First- and family name,
age, function, the event that was photographed, a detailed descrip-
tion for an object
City: In which city the photo was taken
Country: In country the photo was taken
Notes / special instructions: is there a problem with personal rights
of people shown, somebody will need to get glared.Are there spe-
cial claims? But also: Are there technical specialties to the photograph
taken (digital camera, the photo is low key on purpose or with a color
cast or limitations for publication?
Author: Who performed the image description, who can be asked in
case of questions?
Source: enter your Email address here in case of further questions and
if you are sending your photos for the frst time to your client or an
editorial offce.
3.2 IPTC
Image
Engineering
Beyond that there are such things, that do not belong into any IPTC-
felds.
The information about your bank account are not of interest for ex-
ample. Not because you are not supposed to get paid but rather be-
cause in case of a future database query for images from the city of
"Berlin" your image will pop-up just because your bank account may
be from a Bank in Berlin.
This data will be dealt with at a later time. They are not supposed to
be in the photo. The same is true for your address. The only time the
name can be supplemented by the city is if there is reasonable chance
for a mix-up, for example Arthur Miller, Dallas.
Pay attention to the fact that others might be working on a different
computing system. Which means, that umlauts and special characters
(i.e. @, /, , ,) need to be avoided. A point and underscore is appropri-
ate for naming. Filenames should not get longer than 26 characters,
since some databases will simply cut off additional characters.
Any image should have the according fle suffx attached, e.g. .jpg for
JPEG or .tif for Tiff fles and thus it does not matter whether you ar
working on a mac or PC. Especially Windows systems need these fle
endings mandatory, in order to identify the fle correctly.
A discussion on the ECI Mailing list also resulted in the following:
The majority of users prefers names containing the capture date.
The following appendixes will be attached optionally as a prefx or
suffx:
Photographer initials (Image portals!)
Job code
Variations number
internal appendixes, that can be entered as IPTC data.
These appendixes are often useful, if the client does not posses an ac-
cording IPTC-capable software: Highres / Lowres fles, 1st / 2nd choice,
image source, royalty free / licensed, ...
A naming by motive description was completely declined as not ap-
plicable for practical use.
"Client tells the graphic designer please download the image
"blair21.jpg" over at Photopool/ Photoweb/ photofnder ...?
The graphic designer will search for this term and receive 3-6 hits ???
3.3 Description - Naming conventions
Image
Engineering
These kind of problems need to be prevented through clearly unique
naming of fles! (Thanks to Robert Dieth for the abstract)
The Photoshop flebrowser shows EXIF- and IPTC-fles and also edits IPTC
data.
Image
Engineering
While most digital photographers are pretty familiar wit JPEG- and
TIFF formats, there is a hidden talent dozing in the cameras used: The
possibility to save your images as raw-format.
Digital Negative Back to the roots
A digital image saved by the camera as a JPEG or TIFF has gone
through an extensive editing process by the internal camera electron-
ics.The data delivered by the sensor run through several corrections,
e.g. white balance, color, saturation, unsharp masking, if so lossy JPEG
compression. As a photographer you only have limited infuence on all
of these processes (via the camera menu) and you pretty much have to
settle with the editing parameters specifed by the manufacturer . The
infuence of these processes specifed by the manufacturer on the qual-
ity of the result will become obvious whenever cameras from multiple
manufacturers use the same CCD, but the results under equal condi-
tions when capturing vary greatly. The raw data format resembles the
analogue flm, developed and enlarged in your own lab respectively
darkroom and therefore meets the requirements for the term "digital
Negative". The photographer will obtain the possibility back to switch
off the applied automatic processes inside the camera with the aid of
the according converters, in order for him to determine himself, how
the image is supposed to look.
What is raw data
In combination with digital cameras raw refers to a fle format by
which the cameras saves the information captured by the image sen-
sor, directly 1:1 to the according memory card. In order to achieve the
kind of quality that can be produced with this kind of format in an
image editing application, a special conversion is required. Due to the
specifc architecture of the sensor an image will not be digitized the
way it is seen on the display or monitor. A special interpretation and
optimization of the digital material is necessary. This is also the reason
for the term RAW we are dealing with the raw information coming
from the sensor.
Interpolation of color
All "single-shot" digital cameras contain a sensor, which in simplifed
terms consists of many photosensitive sensors arranged in a grid-like
array, that only react to the amount of light falling onto them and
which represents an appropriate tonal- / grey value after capturing
4 Processing data
4.1 Importing and Processing raw data
Image
Engineering
/ analysis (the Foveon chip is the
only exception). All these tonal
values together make the cap-
tured image, yet until now it ac-
tually consists of greyscales only,
that have been captured through
a flter, but do not contain any
real colors yet. In order to cre-
ate a digital color image, a flter
is applied to each photosensitive
image sensor, in one of the three
primary colors red, green or blue.
In order to simulate the special green-sensitivity of the human eye, the
sensors contain 25% of red, 25% of blue and 50% of green flters in a
determined array. This determined array is often referred to as "bayer-
matrix". In order to receive a fully colored digital image with 100%
red, green and blue pixels, the colors have to be interpolated with the
aid of very sophisticated algorithms, which are not available for the
according pixel due to the bayer-matrix. If necessary the surroundings
of e.g. an "non-green" pixels will be taken into account, in order to
defne the actual color of that particular pixel.
Ingredients to the raw fle model
The raw fle model consists of three well-defned elements:
Camera data better known as EXIF data, contains information about
the camera used, the applied capture parameters like exposure set-
tings, f-stop, etc.
image parameter these parameters were set at the time of capture,
but can be edited in the raw converter afterwards, which in turn has
direct infuence on the conversion in the converter respectively on the
resulting image. This is about the applied color mode, white balance,
saturation, tonal range and application of unsharp masking.
Image data non interpolated raw data coming directly from the CCD
or the CMOS of the camera with high bitdepth (most of the times
either 12 or 14-bit).
Advantages of raw data
The advantages of the raw data format when used with the according
raw converters, offer a large freedom in the world of digital images
in order to achieve the best quality possible. In return the raw data
requires a lot more memory than JPEGs and it is rather time consum-
ing to edit them image by image. Thus the raw data format still offers
a lot of advantages:
Larger bitdepth that JPEGs or TIFFs: Instead of reducing them to 8-bit
Above: The Foveon
sensor with the three
sensor layers.
Left: sensor with the
conventional bayer
flter.
Image
Engineering
(256 gradations) the tonal values will e saved with the highest possible
bitdepth accordingly.
Subsequent infnitely adjustable white balance, independently from
the settings inside the camera, usually in the range from 2.000 and
10.000 Kelvin.
Selective contrast of single tonal areas, e.g. via contrast curves wit
direct visual feedback on the monitor. If there is an image high in
contrast the details in the shadows, with usually little or detail can be
lightened with contrast curves, while the brighter parts of the images
remain unchanged.
Continuous adjustment of saturation of color in the raw converter in
order to control respectively optimize the color nuances inside the im-
age
continuous control of digital unsharp masking, precise adjustments
individually adopted to every single image.
No lossy compression as in e.g. the regular JPEG processing
Raw converters
In order to achieve the best possible image quality from the digital
camera used, the raw data format should be used. The raw convert-
er is an application, which creates the well known standard TIFF and
JPEG fles, in order to make the further editing process in all image
editing or layout applications possible. All present raw data formats
are proprietary. Which is also the reason why there are actually two
kinds of raw converters. On the one hand there are the raw converters
supplied by the manufacturer, which can convert only their own raw
data. Secondly there are the so-called "third party " raw converters
which can interpreted and edit the majority of different raw data for-
mats. These groups are separated again into two groups with differ-
ent strategies for the raw conversion:
linear conversion: raw interpolation, exposure correction and white
balance are applied in the application itself; image editing and re-
touching is done in an extra application.
"Full Service"- conversion: all necessary steps of the conversion and
the image editing can be performed in one single application.
The linear conversion is mentioned here actually just for the sake of
completeness, since every raw converters has to master this discipline.
The result of a linear conversion is a dark image, that needs to un-
dergo a specifc tonal range optimization. A detailed description of
the complete list of each individual editing step during the conversion
would go beyond the scope of this guide. Merely a simple enumera-
tion in note form shall be suffcient enough to judge the functional
range of the own converter used:
Image
Engineering
Bayer matrix interpolation (ex-
ception camera raw data from
a foveon chip).
Camera specifc profles and
tonal range optimization ge-
neric camera profles for all
camera formats supported in
order to describe the specifc
characteristics of the camera.
White balance stable and
neutral grey balance
exposure correction correc-
tion of over- or underexposure
noise removal flter (optional)
the sooner in the workfow
the more effcient, in order to prevent a potential amplifcation in the
following steps.
Anti-moire flter - disadvantage from the bayer matrix and especially
important for cameras with a weak anti-aliasing flter
control over brightness, contrast, saturation, global / selective color
correction and last but not least "unsharp masking".
Color management support / implementation of ICC profles
Raw-workfow
After opening a raw fle in the raw converter, the user will receive
an already interpolated, full color preview from the available "bayer
fle", that will then be edited interactively with the various tools, dia-
logues or menu options. With the aid of such familiar and trusted ed-
iting tools, like e.g. the gradation curves, the expected results can be
associated via live preview to the fnal result.
With the next steps global corrections like for example USM sharpen-
ing flter, moir- and/or noise reduction will be applied. Finally set-
tings like color space, color depth, output size and resolution will be
applied. Only then the actual conversion will be started, while at the
same time the primary image data will be interpolated from the bay-
er-matrix of the raw fle, in order to create the digital image in color.
This clearly points out the advantages of editing the fle in a "raw
state", because the full bitdepth available is taken into account as a
basis for the computing, opposing a JPEG or TIFF fle based on 8-bit,
created by the in-camera processing.
Exemplary "high-bit" raw fle processing
1.) Choice of the internal working space in regard to the intended
Comparison of linear to "Full Service" conversion
Image
Engineering
use
2.) Adjust tonal range defne thresholds for highlights and shadows
3.) Adjust gradation - adjust mid tones, for example using numerical
numbers in the histogram or curves like features
4.) Alter white balance if necessary
5.) Color cast removal ideally using color temperature (or an accord-
ing eyedropper tool) as well as a tint / tone slider. An according eye-
dropper tool can be applied simply by clicking inside the image on
neutral grey subject tones. This procedure doesnt always create the
expected result, because quite often the light situation while captur-
ing the image wasnt neutral either.
6.) Activate moire removal
7.) Smoothing and noise removal
8.) Apply unsharpmasking
9.) Choose fnal color bitdepth
10.) Select output and pixel dimensions upsampling on the RAW fle
often gives better results in comparison to already processed fles
11.) Continue editing for example retouching or small enhancements
in the favorite imaging application
In the left image color temperature and tint were set to remove the alleged
color cast; the evening mood was contained but the colors in the sky seem
dull and weak.
The right image is more accordingly to the actual scene after a manual white
balance optimization.
Image
Engineering
Image
Engineering
Different imaging applications offer different ways to optimize alleg-
edly too dark, too weak or too light images. Please avoid in almost
any case the brightness- / contrast slider, since the tonal values will be
changed in a rather "crude" way. The settings of histogram or tonal
range levels respectively gradation curves are much better suited for
this kind of purpose.
Tools for image information
When optimizing images manually it can be mandatory to be able
to follow up on the changes applied. Nearly all imaging applications
offer either a graphical data display in the shape of an histogram for
this kind of purpose or a numerical in the shape of an "info-tool"
or densitometer. Using these two tools at the same time enables the
display and optimization of data at the same time. Additionally both
of these tools make it easier for the user to actually understand the
changes applied.
The densitometer
The densitometer / info-tool displays the actual pixel data ( actually
greyscale values and the related color values of the pixel )of the image
in numerical values in two groups. One group displays the measure-
ments before and the second group displays the measurements after
the optimization of the according greyscale / color values of the mea-
sured image area.
There is one detail that needs to be checked upon which is the radius
of the densitometer / info tool. In general it can be set to measure a
single or an average of a certain number of pixels, e.g. a 3x3 or 5x5
pixel radius. Normally an average of numerous pixels should rather be
read because single pixels could read false values due to noise.
The histogram
A histogram is based on the analysis of all pixels from a single digital
image. The distribution and the amount of pixels from different tonal
levels are presented graphically. On the left hand side is the zero point
at which the darkest (black) pixels are settled, the right end is the
origin of the brightest (white) pixels.
The width of the peaks and valleys in-
dicates which spectrum of gradations
an image contains. The height of an
"data peak" indicates how much data
this certain part of the greyscale spec-
trum contains.
4.2 Tonal range levels and gradiation adjustments
Image
Engineering
Manual image optimization
This chapter will analyze the manual approach of the usual adjust-
ments and optimizations.
Optimization of tonal range
The optimization of tonal range is a procedure to adjust brightness
and contrast. If performed in single color channels, the tonal range
optimization is also used for the compensation of color casts.
The quality of an image will be refected by an either "good" re-
spectively "bad" histogram. A "good" image (see example on the
left) normally uses the complete width of the available tonal values.
A "bad" image on the other hand (see example on the right) will not
use the complete width available or it wont reach to the end of the
scale on both sides of the histogram.
Please remember a graphical histogram display represents the com-
plete range of greyscales. This means, in order to make a "good" his-
togram from a "bad" histogram, for example a fat image, a correc-
tion needs to be performed, which spreads the available pixels over
the complete scale of the available tonal values.
An optimization of tonal values thus needs to be performed if the pix-
els in one channel need to be recomputed because they do not cover
the whole available scale of for example 256 tonal values in an 8-bit
workfow in order to improve this range. This creates an improved
saturation in colors and a higher contrast, which means black will be
darker and white even brighter.
Set highlights and shadows
Before setting highlights and shadows, the brightest and darkest spot
in the image needs to be searched for and analyzed whether they
coincide with the important "content areas" of the image afterwards
we can use the histogram dialogue for a manual optimization of high-
Image
Engineering
lights and shadows with the great advantage of actually seeing the
optimization in a live preview while performing them. The following
example is supposed to clarify how to use the histogram in order to
adjust the highlights and shadows.
This is a clipping of the image used prior to any
optimizations, and as we can see it is desperately
asking for some attention because it is simply
fat.
The empty left area of the sample histogram in-
dicates that there are no pixels available in this
area, representing black. It is essential now to re-
apply the minimum highlights as well as the maximum shadows com-
pletely new.
This is done in most imaging
applications by simply drag-
ging the small triangles on end
of the highlights and shadows
to the just about noteworthy
beginning of the important
image data.
Additionally the overall bright-
ness of the images gets opti-
mized by dragging or moving
the middle triangle, responsible
for the mid tone brightness, un-
til the desired result is achieved.
Note:
As a general rule the use of the brightness /
contrast menu in order to adjust the bright-
ness should be prevented in any case because
it is the only way to preserve the carefully ad-
Image
Engineering
justed highlights and shadow levels. The gradation curves dialogue is
much more appropriate to adjust brightness.
Adjusting gradation
Until now the optimization of tonal values as setting highlights and
shadows were displayed.
The gradation curves dialogue is one of the if not the most powerful
tool to adjust for image optimization as well as optimization of grada-
tion. By this each input value (current state of a pixel) can be assigned
to a continuous output value (specifed state).
The curve in the curves dia-
logue represents the transfor-
mation of the distribution of
the greyscale values between
the original (the original im-
age, lower edge of the grada-
tion square, marked in green)
and the target ( what you
will get, left edge of the gra-
dation square, marked red).
Thus the non manipulated
shape of an gradation curve actually is no curve as the brightness val-
ues of the pixel in an image appear as a straight line in a xy-diagramm
with a 45 degree rise. The x-axis refers to the original brightness (in-
put values) of each pixel, the y-axis represents the new and changed
output values.
Only a straight "curve" thus represents equal input and output values.
The output value of each pixel still represents exactly its input value.
When changing the curves the input and output values will differ.
Adjusting brightness
As describe by now a gradation curves dialogue is much more appro-
priate to adjust the brightness of an image. The input and output
can be changed infnitely simply by adding control points to the curve
or by dragging the existing
points. In order to brighten
up an image, without loosing
the carefully adjusted values
for highlights and shadows,
simply change the shape of
he curve in the middle ton-
al values, the so called mid
tones, in a way for the curve
to bend upwards. By means
Image
Engineering
of the numerical values we can observe pretty well, how the input-
and output values will now differ from each other.
Adjusting contrast
If the overall contrast of an image is adjusted, this will also change the
overall shape of the gradation curve.
An s-shaped curve, that is bend upwards in the highlights section and
fattened downwards in the shadows section, will increase the over-
all contrast of an image. A slight increase of overall contrast pays off
most with most images, since it will give the image a more photo-
graphic look-and-feel.
It may be mentioned to the topics optimizations for tonal values as
well as gradation curves.
Image
Engineering
Even if the reproduction of fne details - thus the resolution - is in
noway related to the number of pixels ( = sample rate), the two terms
are often equated in practice. Digital cameras provide square pixels
in rows and columns. The size of these pixels in the output hereafter
is usually not defned during the exposure. Yet the manufacturers do
save a so called target resolution in the camera, which is usually at 72
pixels per inch or 300 pixels per inch. By this the pixels are defned by
size, which still can be changed later in the subsequent image editing
application. The quality of the image will not be changed by the out-
put size of the pixels. In this case the pixel behaves like the flm grain.
If it remains adequately small or the viewing distance is adequately
big, it will not become visible. Only at a certain size the pixel will be-
come visible - like the flm grain - and can be recognized as such.
For certain applications the pixel size has to be reduced or in excep-
tional cases it has to be enlarged. A reduction is necessary if the im-
ages will be optimized for monitor or web applications. Also when
sending an image via data line it has to be reduced to the required size
in order to save time and costs.
Adding pixels will lead to larger fles, yet will not generate additional
details and therefore should be an exemption (compare 2.3.).
Tip:
For magazines like People or Time images with 4 million pixels (12
mb fles at 8-bit) are suffcient for smaller illustrations. For single or
double page illustrations fles with 18-24 mb will result in a very good
quality in the printed result. The bigger fles can be compressed via
JPEG-format as well (quality level 8 or higher). These images will be-
come approximately 1 mb fles, which will not cause any problems
anymore in times with fast internet or DSL connections.
Images for web pages do not need to be any bigger than 1024 x 768
pixels and should be saved as sRGB images as JPEG-format.
4.3 Adjusting resolution
Image
Engineering
Since different output processes require different degrees of subse-
quent sharpening and the interpolation to a larger as well as to a
smaller fle size has a negative effect on the sharpness, it is recom-
mended to perform the fnal sharpening just before outputting the
image.
An out-of-focus image can not be converted into a sharp image with
the help of sharpening. But a reasonably sharp image will appear a lot
sharper for the human eye by increasing the edge-contrast.
The only adjustable sharpening flter in Photoshop is "Unsharp mask-
ing". The algorithm behind is actually a couple of years old and an
adaption to newer technologies could be useful. Applications like Sil-
verFast or Nik Sharpener are signifcantly better in this aspect.
But the algorithm of the photoshop flter is appropriate enough to
display the functionality of a sharpening flter. The sharpest edge in a
digital image is a black pixel next to a white pixel (magenta line). The
softest edge is a brightness gradient across numerous pixels (yellow
line).
By adjusting the strength of the flter we can defne how vastly the
yellow line will be transferred into the magenta line. The threshold
specifes how far apart the luminance values have to be, in order to be
included by the flter. With a low threshold even the fne differences
like noise will be sharpened. This is not the case with a bigger thresh-
old. Common values for a threshold are usually between 0 and 15. For
a technical photograph the optimum is tending rather towards 0 and
it is signifcantly higher with portraits, in order not to emphasize skin
impurity even more. The preview in photoshop visualizes the result.
The radius specifes how many pixels the flter will affect. It can be
noticed that if the optimum value is above 2 pixels, the basic sharp-
4.4 Additional sharpening
Image
Engineering
ness of the image is not suffcient enough, in order to receive a really
good result. An adjustable blocking of darker areas in the image or
single colors, like for example skin
tones would be desirable. Also some
sort of intermittent control for the
"cross conversion" on the edges
would really improve the flter. This
"cross-conversion" appears if the
sharpening was overdone and will
become visible through white and
black fringing on the edges.
Image
Engineering
Numerous fle formats have been established with digital
photography and each one has been optimized for a special
assignment. One needs to ponder about the intended use or
further processing steps are planned when saving an image.
Some formats are recommended for the use in the internet
but are rather inapplicable when transferring data to a service
provider and the communication with professional prepress.
JPEG
JPEG images have achieved acceptance in digital imaging in those ar-
eas in which small fle sizes are essential despite photorealistic illustra-
tion. It is one of the most important fle formats for digital cameras
and will also be applied whenever electronic images have to be trans-
ferred via internet, ISDN or UMTS (details about JPEG compression un-
der "2.2 fle formats / compression / color depth)
The JPEG-format offers some specialities: It can apply the important
cmyk color space used for prepress, which is necessary for the pro-
fessional optimization of images and the output on a rotary printing
press. Furthermore profles for color management can
be attached to a JPEG fle. Important: IPTC-data can be
integrated into the fle. Image agencies and editorials
manage their photos according to this archiving stan-
dard, which contain e.g. the name of the photographer
and further keywords (see 3.2).
"Progressive JPEG" is recommended for the internet.
Thereby a complete but low resolution version of the
photograph will be displayed frst . Subsequently the
rest of the image will be loaded bit by bit. Another ver-
sion of JPEG is "JPEG 2000", that operates with a different compres-
sion algorithm. The most important imaging applications since are
able to read (and write) it, but is has not prevailed yet.
very appropriate for digital cameras thanks to a strong compression
numerous exchange opportunities between different computers
TIFF
Next to JPEG this format is the most important format for image edi-
tors. Tiff-fles can be coded in 8- or 16-bit (see "fle formats / compres-
sion / color depth") and thus it can manage not only 16,8 million col-
ors yet a couple billion color nuances. This image format is supported
by all major imaging and layout applications. Normally it operates
5. Interchange format for the transfer of data
5.1 Recommended fle formats
Image
Engineering
without image compression, but if offers the opportunity to down-
size the fle via LZW-compression. Tiffs can contain either RGB, CMYK
or LAB data. By this they cover the most important
color spaces / -models within photography and pro-
fessional post processing. Additionally color profles
can be integrated into the TIFF fle. And just like JPEG
IPTC data can be integrated into the TIFF fle in order
to save copyright information. TIFF fle can contain
alpha channels and mapping paths to save transpar-
ent elements in the image and additionally they can
consist of multiple image layers.
-save images data without compression
essential color models, color profles and alpha chan-
nels can be integrated
EPS
EPS was originally developed for displaying vector graphics (logos
etc.) and basically consists of a "data container" encapsulated in a
postscript fle. Which is why it was named "Encapsulated Postscript".
But the EPS-fle can also contain pixel data thus photos whereas
the image data can also be presented as a space saving
JPEG image. EPS fles can be saved either as RGB- or as
CMYK data. Single elements of the image can be saved
as a mapping path, in order to place the photo in front
of a different colored background or in order to align
text around the outline. The EPS fle can contain a low
resolution preview of the photo, which is why it is pref-
erably used in a layout applications. During the layout stage the ap-
plication accesses just that and used the highres data just prior to the
output. The behavior of EPS fles is problematic in relation to color
management. With a trick ICC profles can be attached to the fle by
photoshop while editing the image but generally these wont be ac-
cepted by the layout application and therefore the color in the fnal
print will be a game of luck.
saving without compression possible
different color modes will be checked
integration into layout applications (mapping path etc.)
problematic in ICC workfow
PSD
PSD is the photoshop native fle format. It can save images in RGB and
CMYK as well as other color models, furthermore the unique layers
of an image will be preserved. It is less adequate for the fle transfer
because the recipient implicitly has to open the fles with photoshop.
Image
Engineering
There are other applications that understand this format, but there
may occur problems due to incompatibility. This is also true when ex-
changing between different versions of photoshop.
ideal working format within photoshop
many color models possible, individual layers for composings
will remain in the fles
less appropriate for fle transfer
RAW
The RAW format is not appropriate for exchanging fles either. Since
RAW images are almost always saved in a special proprietary format
by the manufacturer, the recipient of the images will need to have the
according software, in order to open this format ( see also "Photogra-
phy fle formats / color depth / compression").
Raw does offer the advantage that image data can be saved 1:1 from
the CCD, in order to retain extensive possibilities for optimization.
More and more professional photographers use this format, yet in or-
der to transfer the images it has to be converted into one of the com-
mon standard fle formats.
high quality premium alternative to save images in the camera
has to be converted to a standard format with special software when
exchanging fles
GIF
GIF is a graphics format developed for the internet by the online ser-
vice provider CompuServe. The "graphics interchange format" is able
to display 256 colors only and thus can be used to display graphics
(logos, buttons, etc.) with defned spot colors and a strong image com-
pression. Which is why it was established as standard for graphics on
the internet. Inversion GIF89a it can save multiple images within a
single fle, that can be played as an animation.
displaying logos and graphics (buttons, background graphics etc.) on
the internet
animation on the internet
PNG
The PNG (pronounced "ping") was developed in order to by-pass the
disadvantages of the GIF format. The "portable network graphic" for-
mat can display up to 16,8 million colors. Even though all modern in-
ternet browsers can read PNG and the World Wide Web Consortium
W3C formally recommends it, it has not yet been enforced completely.
PNG compresses the images with a lossless algorithm and can contain
one transparency channel.
images with transparencies for the internet and multimedia applica-
Image
Engineering
tions
minor distribution
GIF and PNG have almost no meaning in the area of digital photog-
raphy.
DNG
Adobe Systems has done an innovative step, which has heated many
discussions since. The innovation consisted of giving the users and
manufacturers a unifed fle format for RAW data to hand. On the
frst impression this does not seem to innovative yet it is no reality in
digital photography, because until today almost every manufacturer
of digital cameras saves it`s own RAW format. For the data to remain
processable, the manufacturers offer their own converters to each ap-
propriate format. Yet since with each introduction of a new model
a new RAW format is introduced, it is possible for older formats not
to be supported anymore in 4-8 years by the according tools or other
third-party converters.
DNG offers a unifed RAW format as an open standard, in which a
large bandwidth of camera models can save all necessary image infor-
mation into a unifed fle format. A future-proof solution for archiving
images in DNG format is possible with the also freely available DNG
converter by Adobe Systems. The DNG-converters at the time supports
all RAW formats by the common camera models. The initiators em-
phasize that by using the DNG-format it is guaranteed that a once
archived RAW fle can be opened and processed even in 10 or more
years. Whether the DNG format will become accepted as standard
(thus can still be opened by Photoshop in 10 years or more) strongly
depends on when and how the camera manufacturers will start sup-
porting this fle format. At the moment many users and potential cli-
ents are still waiting for the frst manufacturer to offer native support
for the DNG format.
Note:
If the DNG conversion is supposed to be used
for archiving in RAW format, the option
to preserve the raw fle should be chosen.
By this the image data will be saved in its
original "mosaic format" (see 4.1 What are
RAW fles). If improved methods for conver-
sion should become available in a couple of
years, these could be applied to the archived
RAW fles. In return these could not be ap-
plied to an already linear converted fle any-
more.
Image
Engineering
Open DNG fles in Adobe Photoshop CS:
For opening and displaying DNG-fles in Photoshop CS you will need
the latest Camera Raw 2.3 plug-in. You will receive access to the latest
version via the downloads section on www.adobe.com.
PDF/X3
Early september 2002 the ISO-standard 15930-3:2002 better known
as PDF/X3 was published. PDF/X3 is a standard in the shape of a fle
format. It is based on the PDF-format and describes a standardization
for the transfer of digital lithographs. PDF/X 3 constitutes the concept
of media neutral (device independent) storage and transfer of digital
originals for print (e.g. Offset- or newspaper print), yet accompanied
with an explicit description, for which output process the fle has been
created and optimized for.
The strong focus on process control is actually intentional and because
of it so essential, because all participating parties will beneft from the
high level of security:
the creator of the digital lithographs (graphic-, designer or more of-
ten the photographer themselves) will be granted security that the
fles delivered were transferred correct content-wise as well as color-
wise to post-processing.
The service provider (e.g. printing plant, newspaper publisher) can as-
sume with safety, that the digital originals provided to him will run
smoothly through his internal processes without mistakes. In addition
to that he will be able to achieve (or even exceed (the expected re-
sults) with the originals provided.
In order to provide method for the creation and processing of PDF/X3
to the rookie in particular, two "cooking recipes" were created by the
ECI working group PDF/X3, these explain the important steps in all
major DTP-applications. These "cooking recipes" can be downloaded
in the download section of http://www.eci.org .
While part 1 describes the creation of process-optimized data with
QuarkXpress as layout application, part 2 shows how-to create media
neutral PDF/X3 fles when applying InDesign 2.
Links
www.jpeg.org/
http://en.wikipedia.org/wiki/Encapsulated_Postscripthttp://www.eci.
org/eci/en/032_pdfx3.php
http://www.pdfxreport.com/faq.html
http://www.pdfx3.org/
Image
Engineering
In order for large amounts of data to be delivered safely to the edi-
torial staff the postal service makes sense. Burned on a CD-ROM the
images are saved cost-effective and in large amounts, and if the CD
is packaged correctly, the images will not take harm. Yet short-term
delivery deadlines and speedy corrections often demand for a fast-
er more fexible solution. This is where the internet comes into play,
while the transfer rate sets limits though.
Email attachments
Email transfer is one of the most important alternatives for fle trans-
fer. Any number of fles can be attached to an Email as an Attachment
in order to be sent to the recipient. The transfer of images with rather
big fles is only recommended though if both parties have at least fast
DSL access. Otherwise the transfer will too long. Many mail provider
also limit the permitted size of attachments which may also block the
transfer. For this reason it is recommended to attach the individual
fles to single mails and send off separately. Due to the size-limitations
the JPEG fle format prevailed. Layouts though will be sent as low
resolution PDF-attachments to the Email.
The speed of the fle transfer is dependent on the internet connection
of the sender as well as the recipient: a DSL connection operates with
standard 128 kilobit/s upload- and 768 kilobit/s download speed, mod-
ern solutions allow even higher transfer speeds of 384 kilobit/s upload
and 3.072 kilobit/s download (as of august 2004).
ISDN
The fle transfer via ISDN-connection has been established as standard
for professional prepress. Since Apple Macintoshs are often used in this
area, many editorial and printing plants are
relying on Leonardo-ISDN cards from Herm-
stedt, which have become a de facto stan-
dard. On the PC side the Fritz-cards from
AVM are used more often. In both cases
the hardware in combination with the pro-
vided software takes care of establishing a
connection, the communication and trans-
fer to the recipient. While the differences
between the different operating systems
actually play major role, yet are no insuper-
able barrier anymore. You simply enter the
phone number of the ISDN-remote station
on the graphical interface and subsequent-
5.2 Communication transfer
Image
Engineering
ly choose the folder to be transferred. The native Hermstedt protocol
for fle transfer plays a major part when a communication between
two ISDN-cards is applied, with Fritz cards on the PC side the Euro-fle-
Transfer protocol is applied in most cases.
The transfer speed is 64 kilobit/s per ISDN-channel. Often times the
used software / hardware solutions offer two ISDN channels for fle
transfer, so 128 kilobit/s can be achieved. There are solutions avail-
able in order to connect a number of ISDN-cards to an array in order
to achieve even higher transfer rates. The connection of more than
one of ISDN-channel will create signifcantly higher costs though for
transferring fles.
FTP
The fast File Transfer Protocol has
proven itself for the direct Internet fle
transfer without E-mail. The sender re-
quires access to an Ftp-server, to which
he will save the digital images and text.
In order to gain access to the FTP-serv-
er he requires some details: The target
has to bee addresses by an IP-address
or URL, thus a numerical order like
192.168.150.90 or ftp://ftp.provider.
com. If it is a non-public server a user-
name and password is required addi-
tionally. The only disadvantage is that it is not possible to address the
recipient directly or a search like in a database is not possible, yet the
images reside the images reside on the ftp-server organized in a fle
directory. In terms of speed the same details are true as for E-mails.
Website and database
If images, pdf fles or other data are supposed to be organized or pre-
sented with a preview, then one has to arrange a website. Similar to
the pure FTP-server the images will be provided to a webserver. A link
on a html-page will refer to it, from which it can be downloaded. The
Hypertext Transfer Protocol (http) will regulate the data stream. HTTP
was originally designed for websites, yet in the meantime it regulates
the transfer of any kind of data.
More professional solutions are structured databases, which can be
searched systematically. One of the most popular database systems
is for example mySQL, which receives a graphical interface with the
script language php. Those kind of collections available online allow
for a comfortable search, yet have to be organized constantly by the
user. In terms of speed the same details are true as for E-mails.
Image
Engineering
Direct connection via modem
The direct communication between two modems is actually very rarely
by now. The modem will dial the number from the other person, that
will accept the call. Then by entering commands like "upload" or
"download" to the command line console fles are sent or received
from the download-area. The detour via a provider is unnecessary
herewith, since the transfer takes place directly between the sender
or recipient without the intermediary. Standard modems, often times
already built into PCC or Laptops, will achieve fle transfer rates of up
to 56 kilobit/s.
UMTS
Ever since the introduction of the Multimedia
Message System images or similar can be send
from a mobile phone as well. Even in distant
areas data can be transferred provided that
the device used supports the standard.
The weak point so far is transfer speed. The
present standard GSM sends a maximum of
384 k/bits on 8 channels only if the mobile
phone provider supports special techniques
like Enhanced Global Rates for Global Evolu-
tion (EDGE). UMTS is under development in
germany at the moment, which will support
adequate transfer speed s of 2 Megabit per
second by default. In case the recipient will
have less bandwidth available there will be no
problem since the fles are commonly saved on
the internet. The receiver will informed about
a link, under which the material can be found.
Image
Engineering
The photojournalism has experienced some signifcant changes
through digital photography which are of technical as well as of con-
tent nature. The production workfow was shortened signifcantly
through the use of digital cameras which consequently leads to the
fact that images can be inserted into the layout faster and even at a
later point of time. Sport events are a popular example for this: It is
digital photography and the possibility to send the images directly to
the editor via mobile phone or Internet / UMTS that enables newspa-
pers to publish the results from a late evening event in the newspa-
per next morning. The accompanying advantages were the reason for
digital cameras to be used almost exclusively in day-to-day journalism.
And thanks to the high-resolution SLR cameras used even large format
image sequences for magazines can be realized without a problem.
Images in the area of "portraits" or "editorial" are produced with
digital solutions as well most of the times.
Flexibility while shooting
For the comparably low resolution newspaper print image data has
been suffcient enough for a long time now. Hereby digital cameras
produce data, that needs to be reduced artifcially by the photogra-
pher, in order for the photo to be transferred to the editor as a digital
fle subsequently. Even for large-format magazines that feature high
screen rulings printed on premium high-quality papers there are no
more limitations left. SLR cameras with up to 14 megapixels (as of
August 2004) will provide a fle size for example that can easily realize
a DIN A3 page.
Transportable solutions for medium- or large-for-
mat cameras with sensors and up to 22 megapixel
resolution make even larger image formats pos-
sible, if used in multishot-mode. These will simply
move the CCD-sensors slightly between the single
exposures and create huge fles, that are used
mainly for advertising- or packshot photography.
Serial images and zooms
Speed has increased extremely with digital cameras
as well. Digital SLRs with serial shots of more than 8
images per second and 40 images in a row are avail-
able, whereupon those images are still available in
the maximum resolution offered by the camera.
6. Application specifc considerations
6.1 Photojournalism
Image
Engineering
The so called focal length multiplier is evenly important, which is
caused by the limitations of the cropped feld of view caused by the
smaller flm frame area in comparison to the 35mm format. It actually
amplifes the tele effect of the lens by factor 1,5 or even 1,6, which is
why photographers in the feld of sport journalism can revert to ex-
tremely powerful solutions., that have not been available so far with
analogue technology. The sRGB-settings of the camera used is recom-
mended as standard, if custom profles can not be created for the
camera. The agencies and editorials are accustomed to this color
space and will process the images according to the
parameters from their own printing facility. Apply-
ing other color spaces when transferring the fles
will need to be discussed in advance with the
service provider.
File transfer and communication
The fle transfer will occur either via Inter-
net or by mobile phone, whereupon the
UMTS technology has not distributed well
, yet thanks to higher bandwidth will gain
signifcance quickly. The images will be saved
and transmitted as JPEGs almost exclusively,
because this fle format can achieve a rather
good image quality with relatively small fle
sizes. It will be up to the photographer whether
he will save the images in this compressed format right away
or shoot in raw mode frst. These will need to be converted into a
standard format like Tiff or JPEG with a special piece of software that
came with the camera and a laptop computer.
Image management and accounting
IPTC-information will be embedded into the images, which will be in-
tegrated into the fles via a special imaging application or special soft-
ware. It contains information about the capture date, photographers
name, keywords for indexing and similar data.
The image agencies will administrate these images according to the
IPTC-informations in the image databases and content management
systems, thus they can selectively search images to a certain topic.
Additionally the IPTC-data simplifes the royalty billing between the
agencies, editorials and the photographer ( see also chapter "5.1 In-
terchange format for the transfer of data").
Tempting and dangerous
With the emerge of digital cameras a certain problem in photo-jour-
Image
Engineering
nalism was amplifed, which did exist before, yet received a growing
role with the potential of digital image editing image manipulation.
Any part of a digital image can be manipulated by applications like
photoshop or other imaging applications.
Anything that could be accomplished in the analogue lab, yet required
signifcant background knowledge, great technical skills and a lot of
time, has now become very simple with the computer which is why
almost anybody can accomplish these kind of image manipulations.
Proposed solutions for image manipulation
Attempts by journalists organizations, to explicitly label a manipu-
lated image in the printed version in order to be displaced from the
authentic image, have not yet been achieved successfully.
Part of the problem consists of the fact that the grey zone between
a technical optimization and the manipulation in form and content is
rather large. The same tools that can be used in an imaging applica-
tion to remove dust from an digital image, can be used in order to re-
move a person from an image as well. The manipulation itself doesnt
have to be extensive. Even subtile corrections, like changing the color
of an image, can lead to a completely different message of the image
content wise.
The reader of a newspaper or magazine, who sees a printed image,
thus needs to know that the image does not refect reality. Even the
postcard idyll of the hotel booked upon the travel brochure can be
deceiting because construction cranes or power poles in the back-
ground are removed from the image within seconds. Many people
are actually aware of this fact but they will suppress it subconsciously
and have accepted it for images used in
advertising.
The photojournalist and the editorial
staff too of course carry an extended
responsibility with digital photography.
A simple and far from perfect retouching
in an image editing application with the
"clone tool". The windmill is erased from
the photo within seconds. Certainly this
can be performed just as good on unwan-
ted persons.
Image
Engineering
Quite often the photographer enjoys more liberties in portrait- than
in product photography, since he needs to pay less attention to the
exact reproduction. The exact reproduction of color is the most impor-
tant factor when photographing products for a catalogue, the por-
trait photographers client expects to be pictured as fattering as pos-
sible. Instead of precise and neutral color reproduction, which most
people would regard as too pale faced, rather saturated skin tones
are expected.
But as soon as one starts shooting for magazines, the same guidelines
apply, as described in the chapter about "advertorial photography".
It has to be clearly secured that all output devices in the workfow are
calibrated and secured by embedded color profles.
Additionally the relevant color space plays a major role in this area
too. For example when scanning and converting an image to sRGB
working space, highly saturated colors will appear rather fat in com-
parison to larger working spaces like Ado-
beRGB or ECI RGB. If the digital image will
be transferred to an exposure unit, sRGB
should be chosen as working space, just
in case the service provider does not ap-
ply proper color management. Yet for the
reproduction in a offset printing process it
has to be separated to CMYK.
Equipment
A good portrait is more than just a compre-
hensive face. It is supposed to display the
character of a person, the environment or
the achievements. So besides the passport
photograph in the studio. a portrait is cre-
ated around nature with transportable photo equipment or around
someones working environment. Next to the camera further equip-
ment makes sense. For example a transportable computer provides
plenty of storage in order to transfer the fles quickly via frewire ca-
ble. Alternatively the images can be saved from the memory card to a
convenient ImageTank. Those are transportable memory card readers
with a built-in harddrive, which can save a multiple of the memory
card.
Post processing
If color gradients will get torn up during post processing due to large
steps in tonal range, the limits of digital post production will be dis-
The left image is shown in sRGB color-
space, AdobeRGB was used as color
space on the right image.
6.2 Portraiture photography
Image
Engineering
played. The 256 shades of an 8-bit will be pushed
to their limits especially in the area of large for-
mat. That is why a photographer will prefer raw
fles in the according proprietary raw-format. Even
though not all bits of the available 16-bits color
depth are used, the data describes better tonal
values, however they have to be edited digitally,
because it is lacking white balance for example
(see 4.1).
The digital alternatives to the classic screw-in
flters can also be applied to this raw material,
without limiting the available nuances too much:
fog- and star flters, black- and white conversion or
soft-focus. The screenshot for example display the
classic fog-flter from Nik Color FX, that creates a
diffuse environment.
The digital retouching is anticipated by the cus-
tomers requirements. In this case the digital tools
help to create a favorable digital image. By this
smaller skin impurities can be removed with the
clone-tool, wrinkles are removed with the soft-fo-
cus tool and a healthy tint is created with the help of adjusting lev-
els.
Links:
www.photoshopforphotographers.com/pscs/download/
PSCS_colmanage.pdf
Photos in Rawformat - or with a
16-bit colordepth - contain signi-
fcantly more shades of color than
standard 8-bit JPEGs.
Image
Engineering
Advertising photography involves product-, fashion- and industrial
photography as well as architectural photography. There is only one
limitation to it: the image statement and the composition have to
comply with the advertising concept of the client.
Problems will occur as soon as the images have to proceed through
individual production steps until the fnal catalogue- or newspaper
print. The applied devices each have their own color space (like RGB o
CMYK) and subsequently have an appropriate color profle. The color
reproduction is massively infuenced by this. Additionally some devices
are able to reproduce only a comparably small portion of the original
color space. Yet in a catalogue it has to be guaranteed that a product
will be printed realistically on paper. Therefore all parties in the pro-
duction process have to comply with exact standards.
Workfow in the Studio
The equipment of a advertising photographer covers a digital SLR as
well medium format cameras with digital backs as far as large format
cameras, which can be used not only in the studio but thanks to laptop
computer for architectural photography too. The printable results are
the technical benchmarks. The camera not only has to deliver a high
resolution but also has run clear of other optical aberrations. Require-
ments in regards to image resolution are defned by the customer.
For example 11 megapixel are expected in advertising
photography these days, Yet higher resolutions are not
really a problem anymore for todays digital backs as well
as D-SLRs.
The key parameters will be defned during exposure:
1.) the color temperature of the studio lights has to be
aligned accordingly because otherwise there is no way
to control different light sources with different color
temperature. The aftereffect: A colored gradient on a
face or a nonuniform illumination of a surface. This kind
of fne-tuning is not only important for the design of the
image but also a question of effciency. Constant results
can be achieved on different shooting days. Work gets
done faster because the different devices dont have to
be readjusted.
2.) The cameras color space defnes the mixtures of the
palette, how many colors can be reproduced and wheth-
er they can be converted fawlessly to a different color
space. Normally the photographer has two of them available during
exposure: sRGB or AdobeRGB. In addition there are the proprietary
The complete range of
devices in the studio have
to comply with com-
prehensible guidelines,
in order to receive the
expected results back from
the printer.
6.3 Advertising photography
Image
Engineering
raw formats. As rule of thumb the working space used by the cam-
era shouldnt be changed without urgent cause. With one exemption
which is if a camera specifc color profle is available in order to con-
vert the image to another device independent color space.
The very limited nuances of an image shot in sRGB will not be multi-
plied through conversion anyways. Indeed we could assume that such
a comprehensive color space like CIE-LAB would universally deliver
the best possible results. However the 256 tonal shades also create an
obstacle: the bigger the color space is, the bigger the offset between
them. The consequence is, that the reduced color gradients from sRGB
will appear streaky. Adobe and ECI are much more adequate for this
purpose. But diffculties may occur here too, due to software- and
camera related problems, if one can not reach that color space coming
from an individual camera profle. This also occurs in the area of com-
pact digital cameras with the even smaller sRGB color space.
3.) All devices, camera, Monitor or printer have to be calibrated. This
is the only way to accomplish a correct match in the studio and on
the reference print (proof), which anticipates the fnal result from the
printer. The photographer has to be able to comprehend the complete
steps on this way. In order to keep this process steady device depen-
dent profles will be used. As so called ICC-profles they are compli-
ant to the standards of the International Color Consortium. But one
should not trust these profles eyeless. Instead they have to be double
checked either by yourself or a qualifed service provider.
For example a camera can be calibrated by shooting a test chart in the
studio, a software like the Gretag Macbeth Profle Maker Pro (www.
gretagmacbeth.com) or the GMG CamFlow (www.gmgcolor.com) will
analyze it.
The next output device, the monitor, will be measured by a spectral
photometer and the appropriate profle will be created by a software.
Complete packages, consisting of a measuring device and a software,
are offered by e.g. Gretag MacBeth, BasICColor or X-rite.
The fnal print of a premium proof as a reference requires even more
accuracy. For example the whitepoint of the paper doesnt always
comply with the requirements of the ICC-profle of the printer. There-
fore special proof paper should be used.
With the help of a common, non-profled inkjet printer, one will not
get too far. A profle and the appropriate paper some inkjet printer
will deliver a rather good simulation. Yet when printing from layout
applications the use of a RIP software, that supports color manage-
ment, is recommended. Here we can take the applications from Color-
Gate, Ef or GMG as an example.
These arrangements can adjust an inkjet printer with the appropri-
ate settings in order for it to simulate the offset printing process and
Image
Engineering
which, if it follows the ISO norm for offset (ISO 12647), is described
by the ICC-profle of the Mediastandard print. These are available for
download from the ECI website and should be embedded into the im-
age fle when transferred to the printer.
As far as Monitor and printer are aligned with each other, a soft-proof
can be simulated on the monitor and subsequently a real proof as
reference for the printer will be created.
A control strip for the Mediastandard print, the FOGRA mediawedge
from the "Forschungsgesellschaft Druck" (www.fogra.org), will be
printed along, which will eventually get analyzed on the compliance
of the tolerances with a spectral photometer and the provided refer-
ence data. In order to compare these prints visually, D50 normlight is
mandatory.
Pack in a box for digital transfer
If the digital data is accompanied by a proof, either the CMYK-version
that was used for the proof print should be transferred along with the
RGB-version of the image, or the image should be saved as PDF-X3 and
delivered with a so called "output intention" (see 5.1). If a packshot is
supposed to be printed without background, the mapping path is usu-
ally embedded. With the help of the path tool in photoshop contours
will be saved as a constant selection. In this case the Encapsulated
Postscript format (EPS) is recommended. Those mapping path can be
embedded in Tiff fles also. Frankly not all output devices will recog-
nize and thus they often lead to problems.
Archival
Generally speaking the original unedited material should be saved,
even if a correction seems necessary. It contains the original EXIF-in-
formation which will eventually get lost during further editing. In any
case an image should contain the IPTC-data, whether it will be ar-
chived or transferred to the editor. The entries comply with a database
standard, which is valid in most professional image agencies. All major
powerful Image management software is able to readout these infor-
mations. Additionally an edited RGB version can be saved. One will
spare the repeated color corrections or the creation of complicated
gradation curves.
Links
www.color.org/
Image
Engineering
THE WORK (AS DEFINED BELOW) IS PROVIDED UNDER THE TERMS OF
THIS CREATIVE COMMONS PUBLIC LICENSE ("CCPL" OR "LICENSE").
THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER APPLICA-
BLE LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UN-
DER THIS LICENSE OR COPYRIGHT LAW IS PROHIBITED.
BY EXERCISING ANY RIGHTS TO THE WORK PROVIDED HERE, YOU AC-
CEPT AND AGREE TO BE BOUND BY THE TERMS OF THIS LICENSE. THE
LICENSOR GRANTS YOU THE RIGHTS CONTAINED HERE IN CONSIDE-
RATION OF YOUR ACCEPTANCE OF SUCH TERMS AND CONDITIONS.
1. Defnitions
1. "Collective Work" means a work, such as a periodical issue, an-
thology or encyclopedia, in which the Work in its entirety in unmodi-
fed form, along with a number of other contributions, constituting
separate and independent works in themselves, are assembled into a
collective whole. A work that constitutes a Collective Work will not be
considered a Derivative Work (as defned below) for the purposes of
this License.
2. "Derivative Work" means a work based upon the Work or upon
the Work and other pre-existing works, such as a translation, musical
arrangement, dramatization, fctionalization, motion picture version,
sound recording, art reproduction, abridgment, condensation, or any
other form in which the Work may be recast, transformed, or adap-
ted, except that a work that constitutes a Collective Work will not
be considered a Derivative Work for the purpose of this License. For
the avoidance of doubt, where the Work is a musical composition or
sound recording, the synchronization of the Work in timed-relation
with a moving image ("synching") will be considered a Derivative
Work for the purpose of this License.
3. "Licensor" means the individual or entity that offers the Work
under the terms of this License.
4. "Original Author" means the individual or entity who created the
Work.
5. "Work" means the copyrightable work of authorship offered un-
der the terms of this License.
6. "You" means an individual or entity exercising rights under this
License who has not previously violated the terms of this License with
respect to the Work, or who has received express permission from the
Licensor to exercise rights under this License despite a previous viola-
tion.
7. License
Image
Engineering
2. Fair Use Rights. Nothing in this license is intended to reduce, limit,
or restrict any rights arising from fair use, frst sale or other limitations
on the exclusive rights of the copyright owner under copyright law or
other applicable laws.
3. License Grant. Subject to the terms and conditions of this License,
Licensor hereby grants You a worldwide, royalty-free, non-exclusive,
perpetual (for the duration of the applicable copyright) license to
exercise the rights in the Work as stated below:
1. to reproduce the Work, to incorporate the Work into one or more
Collective Works, and to reproduce the Work as incorporated in the
Collective Works;
2. to create and reproduce Derivative Works;
3. to distribute copies or phonorecords of, display publicly, perform
publicly, and perform publicly by means of a digital audio transmission
the Work including as incorporated in Collective Works;
4. to distribute copies or phonorecords of, display publicly, perform
publicly, and perform publicly by means of a digital audio transmission
Derivative Works.
5. For the avoidance of doubt, where the work is a musical compo-
sition:
1. Performance Royalties Under Blanket Licenses. Licensor waives
the exclusive right to collect, whether individually or via a performance
rights society (e.g. ASCAP, BMI, SESAC), royalties for the public perfor-
mance or public digital performance (e.g. webcast) of the Work.
2. Mechanical Rights and Statutory Royalties. Licensor waives the
exclusive right to collect, whether individually or via a music rights
agency or designated agent (e.g. Harry Fox Agency), royalties for any
phonorecord You create from the Work ("cover version") and distribu-
te, subject to the compulsory license created by 17 USC Section 115 of
the US Copyright Act (or the equivalent in other jurisdictions).
6. Webcasting Rights and Statutory Royalties. For the avoidance
of doubt, where the Work is a sound recording, Licensor waives the
exclusive right to collect, whether individually or via a performance-
rights society (e.g. SoundExchange), royalties for the public digital
performance (e.g. webcast) of the Work, subject to the compulsory
license created by 17 USC Section 114 of the US Copyright Act (or the
equivalent in other jurisdictions).
The above rights may be exercised in all media and formats whether
now known or hereafter devised. The above rights include the right
to make such modifcations as are technically necessary to exercise the
rights in other media and formats. All rights not expressly granted by
Image
Engineering
Licensor are hereby reserved.
4. Restrictions.The license granted in Section 3 above is expressly made
subject to and limited by the following restrictions:
1. You may distribute, publicly display, publicly perform, or publicly
digitally perform the Work only under the terms of this License, and
You must include a copy of, or the Uniform Resource Identifer for, this
License with every copy or phonorecord of the Work You distribute,
publicly display, publicly perform, or publicly digitally perform. You
may not offer or impose any terms on the Work that alter or rest-
rict the terms of this License or the recipients exercise of the rights
granted hereunder. You may not sublicense the Work. You must keep
intact all notices that refer to this License and to the disclaimer of war-
ranties. You may not distribute, publicly display, publicly perform, or
publicly digitally perform the Work with any technological measures
that control access or use of the Work in a manner inconsistent with
the terms of this License Agreement. The above applies to the Work
as incorporated in a Collective Work, but this does not require the
Collective Work apart from the Work itself to be made subject to the
terms of this License. If You create a Collective Work, upon notice from
any Licensor You must, to the extent practicable, remove from the Col-
lective Work any credit as required by clause 4(b), as requested. If You
create a Derivative Work, upon notice from any Licensor You must, to
the extent practicable, remove from the Derivative Work any credit as
required by clause 4(b), as requested.
2. If you distribute, publicly display, publicly perform, or publicly di-
gitally perform the Work or any Derivative Works or Collective Works,
You must keep intact all copyright notices for the Work and provide,
reasonable to the medium or means You are utilizing: (i) the name
of the Original Author (or pseudonym, if applicable) if supplied, and/
or (ii) if the Original Author and/or Licensor designate another party
or parties (e.g. a sponsor institute, publishing entity, journal) for at-
tribution in Licensors copyright notice, terms of service or by other
reasonable means, the name of such party or parties; the title of the
Work if supplied; to the extent reasonably practicable, the Uniform
Resource Identifer, if any, that Licensor specifes to be associated with
the Work, unless such URI does not refer to the copyright notice or
licensing information for the Work; and in the case of a Derivative
Work, a credit identifying the use of the Work in the Derivative Work
(e.g., "French translation of the Work by Original Author," or "Screen-
play based on original Work by Original Author"). Such credit may be
implemented in any reasonable manner; provided, however, that in
the case of a Derivative Work or Collective Work, at a minimum such
Image
Engineering
credit will appear where any other comparable authorship credit ap-
pears and in a manner at least as prominent as such other comparable
authorship credit.
5. Representations, Warranties and Disclaimer
UNLESS OTHERWISE MUTUALLY AGREED TO BY THE PARTIES IN WRI-
TING, LICENSOR OFFERS THE WORK AS-IS AND MAKES NO REPRESEN-
TATIONS OR WARRANTIES OF ANY KIND CONCERNING THE WORK,
EXPRESS, IMPLIED, STATUTORY OR OTHERWISE, INCLUDING, WITHOUT
LIMITATION, WARRANTIES OF TITLE, MERCHANTIBILITY, FITNESS FOR
A PARTICULAR PURPOSE, NONINFRINGEMENT, OR THE ABSENCE OF
LATENT OR OTHER DEFECTS, ACCURACY, OR THE PRESENCE OF AB-
SENCE OF ERRORS, WHETHER OR NOT DISCOVERABLE. SOME JURIS-
DICTIONS DO NOT ALLOW THE EXCLUSION OF IMPLIED WARRANTIES,
SO SUCH EXCLUSION MAY NOT APPLY TO YOU.
6. Limitation on Liability. EXCEPT TO THE EXTENT REQUIRED BY APPLI-
CABLE LAW, IN NO EVENT WILL LICENSOR BE LIABLE TO YOU ON ANY
LEGAL THEORY FOR ANY SPECIAL, INCIDENTAL, CONSEQUENTIAL, PU-
NITIVE OR EXEMPLARY DAMAGES ARISING OUT OF THIS LICENSE OR
THE USE OF THE WORK, EVEN IF LICENSOR HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES.
7. Termination
1. This License and the rights granted hereunder will terminate au-
tomatically upon any breach by You of the terms of this License. Indi-
viduals or entities who have received Derivative Works or Collective
Works from You under this License, however, will not have their li-
censes terminated provided such individuals or entities remain in full
compliance with those licenses. Sections 1, 2, 5, 6, 7, and 8 will survive
any termination of this License.
2. Subject to the above terms and conditions, the license granted
here is perpetual (for the duration of the applicable copyright in the
Work). Notwithstanding the above, Licensor reserves the right to re-
lease the Work under different license terms or to stop distributing
the Work at any time; provided, however that any such election will
not serve to withdraw this License (or any other license that has been,
or is required to be, granted under the terms of this License), and
this License will continue in full force and effect unless terminated as
stated above.
8. Miscellaneous
!0! Compendium lor digilal pholography
Image
Engineering
1. Each time You distribute or publicly digitally perform the Work or
a Collective Work, the Licensor offers to the recipient a license to the
Work on the same terms and conditions as the license granted to You
under this License.
2. Each time You distribute or publicly digitally perform a Derivative
Work, Licensor offers to the recipient a license to the original Work
on the same terms and conditions as the license granted to You under
this License.
3. If any provision of this License is invalid or unenforceable under
applicable law, it shall not affect the validity or enforceability of the
remainder of the terms of this License, and without further action by
the parties to this agreement, such provision shall be reformed to the
minimum extent necessary to make such provision valid and enforce-
able.
4. No term or provision of this License shall be deemed waived and no
breach consented to unless such waiver or consent shall be in writing
and signed by the party to be charged with such waiver or consent.
5. This License constitutes the entire agreement between the parties
with respect to the Work licensed here. There are no understandings,
agreements or representations with respect to the Work not specifed
here. Licensor shall not be bound by any additional provisions that
may appear in any communication from You. This License may not be
modifed without the mutual written agreement of the Licensor and
You.

Digipix3 v300 en

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Digipix3 v300 en

Uploaded by

Copyright:

Available Formats

!

Compendium lor digilal pholography

(*) CIE = Commission Internationale de lEclairage. Defned in 1931 the

You might also like